diff --git a/.github/TEAM_MEMBERS b/.github/TEAM_MEMBERS
index e5f8f000e0..a662c7c063 100644
--- a/.github/TEAM_MEMBERS
+++ b/.github/TEAM_MEMBERS
@@ -13,3 +13,4 @@ R44VC0RP
 rekram1-node
 thdxr
 simonklee
+vimtor
diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
index 2ae3fc6f2f..e1a62ae9ca 100644
--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@@ -73,7 +73,7 @@ Replace `<platform>` with your platform (e.g., `darwin-arm64`, `linux-x64`).
   - `packages/opencode`: OpenCode core business logic & server.
   - `packages/opencode/src/cli/cmd/tui/`: The TUI code, written in SolidJS with [opentui](https://github.com/sst/opentui)
   - `packages/app`: The shared web UI components, written in SolidJS
-  - `packages/desktop`: The native desktop app, built with Tauri (wraps `packages/app`)
+  - `packages/desktop`: The native desktop app, built with Electron (wraps `packages/app`)
   - `packages/plugin`: Source for `@opencode-ai/plugin`
 
 ### Understanding bun dev vs opencode
@@ -123,33 +123,21 @@ This starts a local dev server at http://localhost:5173 (or similar port shown i
 
 ### Running the Desktop App
 
-The desktop app is a native Tauri application that wraps the web UI.
+The desktop app is an Electron application that wraps the web UI.
 
-To run the native desktop app:
-
-```bash
-bun run --cwd packages/desktop tauri dev
-```
-
-This starts the web dev server on http://localhost:1420 and opens the native window.
-
-If you only want the web dev server (no native shell):
+To run the desktop app in development:
 
 ```bash
 bun run --cwd packages/desktop dev
 ```
 
-To create a production `dist/` and build the native app bundle:
+To create a production build and package the app:
 
 ```bash
-bun run --cwd packages/desktop tauri build
+bun run --cwd packages/desktop build
+bun run --cwd packages/desktop package
 ```
 
-This runs `bun run --cwd packages/desktop build` automatically via Tauri’s `beforeBuildCommand`.
-
-> [!NOTE]
-> Running the desktop app requires additional Tauri dependencies (Rust toolchain, platform-specific libraries). See the [Tauri prerequisites](https://v2.tauri.app/start/prerequisites/) for setup instructions.
-
 > [!NOTE]
 > If you make changes to the API or SDK (e.g. `packages/opencode/src/server/server.ts`), run `./script/generate.ts` to regenerate the SDK and related files.
 
diff --git a/UPSTREAM.md b/UPSTREAM.md
index 653ce2042e..75989bc54d 100644
--- a/UPSTREAM.md
+++ b/UPSTREAM.md
@@ -77,6 +77,8 @@ Each upstream has its own append-only table. Add a row every time you pull.
 | 2026-05-01 | `af3998c8a` | `21f8027ef` | bcode | Merged upstream release point for v1.14.31 (`sync release versions for v1.14.31` on `dev`). 212 upstream commits across v1.14.27–v1.14.31. Conflicts: `.github/workflows/{deploy,publish}.yml` (kept our deletions per PR #14), `bun.lock` (regenerated), `packages/opencode/package.json` (kept name, bumped to 1.14.31), `packages/opencode/src/agent/agent.ts` (kept browser-sessions whitelist + took upstream's new `Global.Path.tmp` whitelist addition — both go in the same `whitelistedDirs` array), `packages/opencode/src/config/config.ts` (kept `bcode.json/bcode.jsonc` filenames + `bcode.sh` config schema URL; adopted upstream's `mergeConfig` helper pattern, retiring `mergeDeep(pipe(...))` chain), `packages/opencode/src/session/session.ts` (kept `.bcode/plans` rename; adopted upstream's new `(input, instance: InstanceContext)` signature using `instance.project`/`instance.worktree`), `packages/opencode/src/installation/index.ts` (substantial restructure — upstream switched from explicit `Service.of({...})` to `result: Interface = {...}` pattern with self-referential method calls; took upstream verbatim as the base, then re-applied 5 BrowserCode divergences: USER_AGENT prefix, `https://bcode.sh/install` URL, `.bcode/bin` execPath check, BCODE_UPGRADE_DISABLED const, early-return guards in `latest`/`upgrade`/`info`). Yellow-zone audit (7 files: `cli/cmd/tui/app.tsx`, `agent.ts`, `config.ts`, `installation/index.ts`, `session.ts`, `index.ts`, `core/src/global.ts`): customizations preserved (`scriptName("bcode")`, banner, USER_AGENT, `bcode.sh`, `.bcode` paths, `app = "bcode"`). Filtered typecheck: 5/5 passed. PR #29 (v1.14.25) supersedes — close in favor of this PR which covers the same window plus three additional release points. |
 | 2026-05-06 | `21f8027ef` | `773078e81` | bcode | Merged upstream release point for v1.14.39 (`sync release versions for v1.14.39` on `dev`). 178 upstream commits across v1.14.32–v1.14.39. **Notable upstream change pulled in:** PR #20039 — `tool/bash.{ts,txt}` renamed to `tool/shell.{ts,txt}` with sub-folder `tool/shell/{id,prompt}.ts`; `Match` predicate switched from string literal `"bash"` to `ShellID.ToolID`. Auto-merge handled the rename inside our session/index.tsx; only the import block needed manual resolution (kept our `BrowserExecuteTool` import, took upstream's `ShellTool` + `ShellID` imports). **CLI Effect-ification:** `cli/cmd/{run,serve,web,agent}.ts` switched from `async (args) => {...}` handlers to `Effect.fn("Cli.<name>")(function* (args) {...})` form (also adds new `instance: false` and `directory: ...` fields to delegate context loading to the framework). Resolved by adopting upstream handler shape verbatim and re-applying our `bcode` brand strings in `describe:` fields. `cli/cmd/agent.ts` also dropped a level of indentation (legacy nested-effect block) and switched from `Instance.worktree`/`AppRuntime.runPromise(Agent.Service.use(...))` to `ctx.worktree`/`Effect.runPromise(agentSvc.generate(...))`; took upstream wholesale, flipped post-merge `.opencode` → `.bcode` (one site). Conflicts: `.github/workflows/publish.yml` (re-deleted per PR #14), `bun.lock` (regenerated), `README.md` (kept our concise replacement, dropped upstream's reintroduced FAQ/Discord block), `packages/opencode/package.json` (kept name, bumped to 1.14.39), `cli/cmd/{agent,run,serve,web}.ts` (Effect-ification + brand strings), `cli/cmd/tui/routes/session/index.tsx` (Bash→Shell import block). Yellow-zone audit (9 files touched upstream: `core/src/global.ts`, `script/build.ts`, `cli/cmd/tui/app.tsx`, `cli/cmd/tui/routes/session/index.tsx`, `config/config.ts`, `plugin/index.ts`, `provider/provider.ts`, `session/session.ts`, `storage/db.ts`): customizations preserved (`app = "bcode"`, banner+title BrowserCode, BC-prefixed session title, `bcode.json/bcode.jsonc/.bcode/`, `bcode.sh` schema URLs, `bcode.db`, `.bcode/plans`, `bcode-<target>` asset name, BrowserExecute renderer, attribution headers `https://bcode.sh/`/`X-Title: bcode`). Filtered typecheck: 5/5 passed in 13.8s. |
 
+| 2026-05-11 | `773078e81` | `fe594693a` | bcode | Merged upstream release point for v1.14.41 (`sync release versions for v1.14.41` on `dev`). 70 upstream commits across v1.14.40–v1.14.41. **Targeted v1.14.41 instead of latest (v1.14.48, 289 commits)** because v1.14.42 lands three sweeping refactors that warrant separate review: PR #24712 native LLM core foundation, PR #24149 scout agent for repo research, OPENCODE_EXPERIMENTAL_WORKSPACES routing changes (workspace fence headers, fixed-id routing, claim detached sessions). Splitting keeps each sync mechanical. **Notable upstream change pulled in:** PR #26054 `well-known/opencode` remote config now supports an external `remote_config` URL with substitution + headers, fetched and merged via the new `substituteWellKnownRemoteConfig` + `mergeConfig` helpers in `config/config.ts`; adopted upstream's new fetch-and-merge body verbatim, kept our `bcode.sh/config.json` default `$schema`. **Agent dir rename `agent/` → `agents/`:** upstream renamed `.opencode/agent/` to `.opencode/agents/` (PR #14427); adopted as `.bcode/agents/` for consistency. Updated `cli/cmd/agent.ts` create-target path and `feature-plugins/home/tips-view.tsx` agent-tip string. Note: `config/agent.ts:129` already accepted both `agent/` and `agents/` glob patterns, so existing `.bcode/agent/` setups keep working. Conflicts: 22 README translation files (`README.{ar,bn,br,bs,da,de,es,fr,gr,it,ja,ko,no,pl,ru,th,tr,uk,vi,zh,zht}.md`) modified by upstream + deleted by us — re-deleted, all 21 README translations purged at fork rebrand stay purged. `.github/workflows/deploy.yml` (re-deleted per PR #14). `bun.lock` (regenerated). `README.md` (kept our concise BrowserCode replacement, dropped upstream's reintroduced desktop-download table — we don't ship a desktop). `packages/opencode/package.json` (kept name, bumped to 1.14.41). `cli/cmd/agent.ts` (took upstream's `.opencode/agents` rename, flipped to `.bcode/agents`). `config/config.ts` (adopted new remote_config logic, kept `bcode.sh` schema URL). Yellow-zone audit (11 files touched upstream: `core/src/global.ts`, `agent.ts`, `cli/cmd/{agent,run,serve,web}.ts`, `cli/cmd/tui/{app.tsx,routes/session/index.tsx}`, `config/config.ts`, `provider/provider.ts`, `session/session.ts`): customizations preserved (`app = "bcode"`, banner+title BC, BrowserCode GitHub link, `bcode.sh` HTTP-Referer/X-Title/X-Source across 8 providers, Cerebras `X-Cerebras-3rd-Party-Integration: bcode`, `.bcode/plans`, `Skills` import, BrowserExecute renderer). Filtered typecheck: 6/6 passed in 8.7s. |
+
 ### browser-use/browser-harness → `packages/bcode-browser/harness/`
 
 **Upstream:** https://github.com/browser-use/browser-harness
diff --git a/bun.lock b/bun.lock
index ee6a1baf95..15d332f0b1 100644
--- a/bun.lock
+++ b/bun.lock
@@ -29,7 +29,7 @@
     },
     "packages/app": {
       "name": "@opencode-ai/app",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "dependencies": {
         "@kobalte/core": "catalog:",
         "@opencode-ai/core": "workspace:*",
@@ -114,7 +114,7 @@
     },
     "packages/console/app": {
       "name": "@opencode-ai/console-app",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "dependencies": {
         "@cloudflare/vite-plugin": "1.15.2",
         "@ibm/plex": "6.4.1",
@@ -148,7 +148,7 @@
     },
     "packages/console/core": {
       "name": "@opencode-ai/console-core",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "dependencies": {
         "@aws-sdk/client-sts": "3.782.0",
         "@jsx-email/render": "1.1.1",
@@ -175,7 +175,7 @@
     },
     "packages/console/function": {
       "name": "@opencode-ai/console-function",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "dependencies": {
         "@ai-sdk/anthropic": "3.0.64",
         "@ai-sdk/openai": "3.0.48",
@@ -199,7 +199,7 @@
     },
     "packages/console/mail": {
       "name": "@opencode-ai/console-mail",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "dependencies": {
         "@jsx-email/all": "2.2.3",
         "@jsx-email/cli": "1.4.3",
@@ -223,7 +223,7 @@
     },
     "packages/core": {
       "name": "@opencode-ai/core",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "dependencies": {
         "@effect/opentelemetry": "catalog:",
         "@effect/platform-node": "catalog:",
@@ -254,7 +254,7 @@
     },
     "packages/desktop": {
       "name": "@opencode-ai/desktop",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "dependencies": {
         "drizzle-orm": "catalog:",
         "effect": "catalog:",
@@ -296,11 +296,19 @@
         "@lydell/node-pty-linux-x64": "1.2.0-beta.10",
         "@lydell/node-pty-win32-arm64": "1.2.0-beta.10",
         "@lydell/node-pty-win32-x64": "1.2.0-beta.10",
+        "@parcel/watcher-darwin-arm64": "2.5.1",
+        "@parcel/watcher-darwin-x64": "2.5.1",
+        "@parcel/watcher-linux-arm64-glibc": "2.5.1",
+        "@parcel/watcher-linux-arm64-musl": "2.5.1",
+        "@parcel/watcher-linux-x64-glibc": "2.5.1",
+        "@parcel/watcher-linux-x64-musl": "2.5.1",
+        "@parcel/watcher-win32-arm64": "2.5.1",
+        "@parcel/watcher-win32-x64": "2.5.1",
       },
     },
     "packages/enterprise": {
       "name": "@opencode-ai/enterprise",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "dependencies": {
         "@opencode-ai/core": "workspace:*",
         "@opencode-ai/ui": "workspace:*",
@@ -329,7 +337,7 @@
     },
     "packages/function": {
       "name": "@opencode-ai/function",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "dependencies": {
         "@octokit/auth-app": "8.0.1",
         "@octokit/rest": "catalog:",
@@ -345,14 +353,14 @@
     },
     "packages/opencode": {
       "name": "@browser-use/browsercode-core",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "bin": {
         "bcode": "./bin/bcode",
       },
       "dependencies": {
         "@actions/core": "1.11.1",
         "@actions/github": "6.0.1",
-        "@agentclientprotocol/sdk": "0.16.1",
+        "@agentclientprotocol/sdk": "0.21.0",
         "@ai-sdk/alibaba": "1.0.17",
         "@ai-sdk/amazon-bedrock": "4.0.96",
         "@ai-sdk/anthropic": "3.0.71",
@@ -489,7 +497,7 @@
     },
     "packages/plugin": {
       "name": "@opencode-ai/plugin",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "dependencies": {
         "@opencode-ai/sdk": "workspace:*",
         "effect": "catalog:",
@@ -524,7 +532,7 @@
     },
     "packages/sdk/js": {
       "name": "@opencode-ai/sdk",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "dependencies": {
         "cross-spawn": "catalog:",
       },
@@ -539,7 +547,7 @@
     },
     "packages/slack": {
       "name": "@opencode-ai/slack",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "dependencies": {
         "@opencode-ai/sdk": "workspace:*",
         "@slack/bolt": "^3.17.1",
@@ -574,7 +582,7 @@
     },
     "packages/ui": {
       "name": "@opencode-ai/ui",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "dependencies": {
         "@kobalte/core": "catalog:",
         "@opencode-ai/core": "workspace:*",
@@ -623,7 +631,7 @@
     },
     "packages/web": {
       "name": "@opencode-ai/web",
-      "version": "1.14.39",
+      "version": "1.14.41",
       "dependencies": {
         "@astrojs/cloudflare": "12.6.3",
         "@astrojs/markdown-remark": "6.3.1",
@@ -747,7 +755,7 @@
 
     "@adobe/css-tools": ["@adobe/css-tools@4.4.4", "", {}, "sha512-Elp+iwUx5rN5+Y8xLt5/GRoG20WGoDCQ/1Fb+1LiGtvwbDavuSk0jhD/eZdckHAuzcDzccnkv+rEjyWfRx18gg=="],
 
-    "@agentclientprotocol/sdk": ["@agentclientprotocol/sdk@0.16.1", "", { "peerDependencies": { "zod": "^3.25.0 || ^4.0.0" } }, "sha512-1ad+Sc/0sCtZGHthxxvgEUo5Wsbw16I+aF+YwdiLnPwkZG8KAGUEAPK6LM6Pf69lCyJPt1Aomk1d+8oE3C4ZEw=="],
+    "@agentclientprotocol/sdk": ["@agentclientprotocol/sdk@0.21.0", "", { "peerDependencies": { "zod": "^3.25.0 || ^4.0.0" } }, "sha512-ONj+Q8qOdNQp5XbH5jnMwzT9IKZJsSN0p0lkceS4GtUtNOPVLpNzSS8gqQdGMKfBvA0ESbkL8BTaSN1Rc9miEw=="],
 
     "@ai-sdk/alibaba": ["@ai-sdk/alibaba@1.0.17", "", { "dependencies": { "@ai-sdk/openai-compatible": "2.0.41", "@ai-sdk/provider": "3.0.8", "@ai-sdk/provider-utils": "4.0.23" }, "peerDependencies": { "zod": "^3.25.76 || ^4.1.8" } }, "sha512-ZbE+U5bWz2JBc5DERLowx5+TKbjGBE93LqKZAWvuEn7HOSQMraxFMZuc0ST335QZJAyfBOzh7m1mPQ+y7EaaoA=="],
 
@@ -1217,7 +1225,7 @@
 
     "@grpc/grpc-js": ["@grpc/grpc-js@1.14.3", "", { "dependencies": { "@grpc/proto-loader": "^0.8.0", "@js-sdsl/ordered-map": "^4.4.2" } }, "sha512-Iq8QQQ/7X3Sac15oB6p0FmUg/klxQvXLeileoqrTRGJYLV+/9tubbr9ipz0GKHjmXVsgFPo/+W+2cA8eNcR+XA=="],
 
-    "@grpc/proto-loader": ["@grpc/proto-loader@0.8.0", "", { "dependencies": { "lodash.camelcase": "^4.3.0", "long": "^5.0.0", "protobufjs": "^7.5.3", "yargs": "^17.7.2" }, "bin": { "proto-loader-gen-types": "build/bin/proto-loader-gen-types.js" } }, "sha512-rc1hOQtjIWGxcxpb9aHAfLpIctjEnsDehj0DAiVfBlmT84uvR0uUtN2hEi/ecvWVjXUGf5qPF4qEgiLOx1YIMQ=="],
+    "@grpc/proto-loader": ["@grpc/proto-loader@0.8.1", "", { "dependencies": { "lodash.camelcase": "^4.3.0", "long": "^5.0.0", "protobufjs": "^7.5.5", "yargs": "^17.7.2" }, "bin": { "proto-loader-gen-types": "build/bin/proto-loader-gen-types.js" } }, "sha512-wtF6h+DY6M3YaDBPAmvuuA6jV8Sif9MjtOI5euKFWRgCDl5PeDpPsHR9u2l6St5ceY8AZgoNDww5+HvEsXFsGg=="],
 
     "@happy-dom/global-registrator": ["@happy-dom/global-registrator@20.0.11", "", { "dependencies": { "@types/node": "^20.0.0", "happy-dom": "^20.0.11" } }, "sha512-GqNqiShBT/lzkHTMC/slKBrvN0DsD4Di8ssBk4aDaVgEn+2WMzE6DXxq701ndSXj7/0cJ8mNT71pM7Bnrr6JRw=="],
 
@@ -1863,7 +1871,7 @@
 
     "@protobufjs/base64": ["@protobufjs/base64@1.1.2", "", {}, "sha512-AZkcAA5vnN/v4PDqKyMR5lx7hZttPDgClv83E//FMNhR2TMcLUhfRUBHCmSl0oi9zMgDDqRUJkSxO3wm85+XLg=="],
 
-    "@protobufjs/codegen": ["@protobufjs/codegen@2.0.4", "", {}, "sha512-YyFaikqM5sH0ziFZCN3xDC7zeGaB/d0IUb9CATugHWbd1FRFwWwt4ld4OYMPWu5a3Xe01mGAULCdqhMlPl29Jg=="],
+    "@protobufjs/codegen": ["@protobufjs/codegen@2.0.5", "", {}, "sha512-zgXFLzW3Ap33e6d0Wlj4MGIm6Ce8O89n/apUaGNB/jx+hw+ruWEp7EwGUshdLKVRCxZW12fp9r40E1mQrf/34g=="],
 
     "@protobufjs/eventemitter": ["@protobufjs/eventemitter@1.1.0", "", {}, "sha512-j9ednRT81vYJ9OfVuXG6ERSTdEL1xVsNgqpkxMsbIabzSo3goCjDIveeGv5d03om39ML71RdmrGNjG5SReBP/Q=="],
 
@@ -1871,13 +1879,13 @@
 
     "@protobufjs/float": ["@protobufjs/float@1.0.2", "", {}, "sha512-Ddb+kVXlXst9d+R9PfTIxh1EdNkgoRe5tOX6t01f1lYWOvJnSPDBlG241QLzcyPdoNTsblLUdujGSE4RzrTZGQ=="],
 
-    "@protobufjs/inquire": ["@protobufjs/inquire@1.1.0", "", {}, "sha512-kdSefcPdruJiFMVSbn801t4vFK7KB/5gd2fYvrxhuJYg8ILrmn9SKSX2tZdV6V+ksulWqS7aXjBcRXl3wHoD9Q=="],
+    "@protobufjs/inquire": ["@protobufjs/inquire@1.1.1", "", {}, "sha512-mnzgDV26ueAvk7rsbt9L7bE0SuAoqyuys/sMMrmVcN5x9VsxpcG3rqAUSgDyLp0UZlmNfIbQ4fHfCtreVBk8Ew=="],
 
     "@protobufjs/path": ["@protobufjs/path@1.1.2", "", {}, "sha512-6JOcJ5Tm08dOHAbdR3GrvP+yUUfkjG5ePsHYczMFLq3ZmMkAD98cDgcT2iA1lJ9NVwFd4tH/iSSoe44YWkltEA=="],
 
     "@protobufjs/pool": ["@protobufjs/pool@1.1.0", "", {}, "sha512-0kELaGSIDBKvcgS4zkjz1PeddatrjYcmMWOlAuAPwAeccUrPHdUqo/J6LiymHHEiJT5NrF1UVwxY14f+fy4WQw=="],
 
-    "@protobufjs/utf8": ["@protobufjs/utf8@1.1.0", "", {}, "sha512-Vvn3zZrhQZkkBE8LSuW3em98c0FwgO4nxzv6OdSxPKJIEKY2bGbHn+mhGIPerzI4twdxaP8/0+06HBpwf345Lw=="],
+    "@protobufjs/utf8": ["@protobufjs/utf8@1.1.1", "", {}, "sha512-oOAWABowe8EAbMyWKM0tYDKi8Yaox52D+HWZhAIJqQXbqe0xI/GV7FhLWqlEKreMkfDjshR5FKgi3mnle0h6Eg=="],
 
     "@radix-ui/colors": ["@radix-ui/colors@1.0.1", "", {}, "sha512-xySw8f0ZVsAEP+e7iLl3EvcBXX7gsIlC1Zso/sPBW9gIWerBTgz6axrjU+MZ39wD+WFi5h5zdWpsg3+hwt2Qsg=="],
 
@@ -4359,7 +4367,7 @@
 
     "proto-list": ["proto-list@1.2.4", "", {}, "sha512-vtK/94akxsTMhe0/cbfpR+syPuszcuwhqVjJq26CuNDgFGj682oRBXOP5MJpv2r7JtE8MsiepGIqvvOTBwn2vA=="],
 
-    "protobufjs": ["protobufjs@7.5.4", "", { "dependencies": { "@protobufjs/aspromise": "^1.1.2", "@protobufjs/base64": "^1.1.2", "@protobufjs/codegen": "^2.0.4", "@protobufjs/eventemitter": "^1.1.0", "@protobufjs/fetch": "^1.1.0", "@protobufjs/float": "^1.0.2", "@protobufjs/inquire": "^1.1.0", "@protobufjs/path": "^1.1.2", "@protobufjs/pool": "^1.1.0", "@protobufjs/utf8": "^1.1.0", "@types/node": ">=13.7.0", "long": "^5.0.0" } }, "sha512-CvexbZtbov6jW2eXAvLukXjXUW1TzFaivC46BpWc/3BpcCysb5Vffu+B3XHMm8lVEuy2Mm4XGex8hBSg1yapPg=="],
+    "protobufjs": ["protobufjs@7.5.7", "", { "dependencies": { "@protobufjs/aspromise": "^1.1.2", "@protobufjs/base64": "^1.1.2", "@protobufjs/codegen": "^2.0.5", "@protobufjs/eventemitter": "^1.1.0", "@protobufjs/fetch": "^1.1.0", "@protobufjs/float": "^1.0.2", "@protobufjs/inquire": "^1.1.1", "@protobufjs/path": "^1.1.2", "@protobufjs/pool": "^1.1.0", "@protobufjs/utf8": "^1.1.1", "@types/node": ">=13.7.0", "long": "^5.0.0" } }, "sha512-NGnrxS/nLKUo5nkbVQxlC71sB4hdfImdYIbFeSCidxtwATx0AHRPcANSLd0q5Bb2BkoSWo2iisQhGg5/r+ihbA=="],
 
     "proxy-addr": ["proxy-addr@2.0.7", "", { "dependencies": { "forwarded": "0.2.0", "ipaddr.js": "1.9.1" } }, "sha512-llQsMLSUDUPT44jdrU/O37qlnifitDP+ZwrmmZcoSKyLKvtZxpyV0n2/bD/N4tBAAZ/gJEdZU7KMraoK1+XYAg=="],
 
@@ -5639,6 +5647,8 @@
 
     "@opencode-ai/web/@shikijs/transformers": ["@shikijs/transformers@3.20.0", "", { "dependencies": { "@shikijs/core": "3.20.0", "@shikijs/types": "3.20.0" } }, "sha512-PrHHMRr3Q5W1qB/42kJW6laqFyWdhrPF2hNR9qjOm1xcSiAO3hAHo7HaVyHE6pMyevmy3i51O8kuGGXC78uK3g=="],
 
+    "@opentelemetry/otlp-transformer/protobufjs": ["protobufjs@7.5.4", "", { "dependencies": { "@protobufjs/aspromise": "^1.1.2", "@protobufjs/base64": "^1.1.2", "@protobufjs/codegen": "^2.0.4", "@protobufjs/eventemitter": "^1.1.0", "@protobufjs/fetch": "^1.1.0", "@protobufjs/float": "^1.0.2", "@protobufjs/inquire": "^1.1.0", "@protobufjs/path": "^1.1.2", "@protobufjs/pool": "^1.1.0", "@protobufjs/utf8": "^1.1.0", "@types/node": ">=13.7.0", "long": "^5.0.0" } }, "sha512-CvexbZtbov6jW2eXAvLukXjXUW1TzFaivC46BpWc/3BpcCysb5Vffu+B3XHMm8lVEuy2Mm4XGex8hBSg1yapPg=="],
+
     "@opentui/core/diff": ["diff@9.0.0", "", {}, "sha512-svtcdpS8CgJyqAjEQIXdb3OjhFVVYjzGAPO8WGCmRbrml64SPw/jJD4GoE98aR7r25A0XcgrK3F02yw9R/vhQw=="],
 
     "@opentui/solid/@babel/core": ["@babel/core@7.28.0", "", { "dependencies": { "@ampproject/remapping": "^2.2.0", "@babel/code-frame": "^7.27.1", "@babel/generator": "^7.28.0", "@babel/helper-compilation-targets": "^7.27.2", "@babel/helper-module-transforms": "^7.27.3", "@babel/helpers": "^7.27.6", "@babel/parser": "^7.28.0", "@babel/template": "^7.27.2", "@babel/traverse": "^7.28.0", "@babel/types": "^7.28.0", "convert-source-map": "^2.0.0", "debug": "^4.1.0", "gensync": "^1.0.0-beta.2", "json5": "^2.2.3", "semver": "^6.3.1" } }, "sha512-UlLAnTPrFdNGoFtbSXwcGFQBtQZJCNjaN6hQNP3UPvuNXT1i82N26KL3dZeIpNalWywr9IuQuncaAfUaS1g6sQ=="],
@@ -5655,6 +5665,8 @@
 
     "@protobuf-ts/plugin/typescript": ["typescript@3.9.10", "", { "bin": { "tsc": "bin/tsc", "tsserver": "bin/tsserver" } }, "sha512-w6fIxVE/H1PkLKcCPsFqKE7Kv7QUwhU8qQY2MueZXWx5cPZdwFupLgKK3vntcK98BtNHZtAF4LA/yl2a7k8R6Q=="],
 
+    "@protobufjs/fetch/@protobufjs/inquire": ["@protobufjs/inquire@1.1.0", "", {}, "sha512-kdSefcPdruJiFMVSbn801t4vFK7KB/5gd2fYvrxhuJYg8ILrmn9SKSX2tZdV6V+ksulWqS7aXjBcRXl3wHoD9Q=="],
+
     "@rollup/pluginutils/estree-walker": ["estree-walker@2.0.2", "", {}, "sha512-Rfkk/Mp/DL7JVje3u18FxFujQlTNR2q6QfMSMB7AvCBx91NGj/ba3kCfza0f6dVDbw7YlRf/nDrn7pQrCCyQ/w=="],
 
     "@sentry/bundler-plugin-core/glob": ["glob@9.3.5", "", { "dependencies": { "fs.realpath": "^1.0.0", "minimatch": "^8.0.2", "minipass": "^4.2.4", "path-scurry": "^1.6.1" } }, "sha512-e1LleDykUz2Iu+MTYdkSsuWX8lvAjAcs0Xef0lNIu0S2wOAzuTxCJtcd9S3cijlwYF18EsU3rzb8jPVobxDh9Q=="],
@@ -6129,8 +6141,6 @@
 
     "proper-lockfile/signal-exit": ["signal-exit@3.0.7", "", {}, "sha512-wnD2ZE+l+SPC/uoS0vXeE9L1+0wuaMqKlfz9AMUo38JsyLSBWSFcHR1Rri62LZc12vLr1gb3jl7iwQhgwpAbGQ=="],
 
-    "protobufjs/@types/node": ["@types/node@22.13.9", "", { "dependencies": { "undici-types": "~6.20.0" } }, "sha512-acBjXdRJ3A6Pb3tqnw9HZmyR3Fiol3aGxRCK1x3d+6CDAMjl7I649wpSd+yNURCjbOUGu9tqtLKnTGxmK6CyGw=="],
-
     "proxy-addr/ipaddr.js": ["ipaddr.js@1.9.1", "", {}, "sha512-0KI/607xoxSToH7GjN1FfSbLoU0+btTicjsQSWQlh/hZykN8KpmMf7uYwPW3R+akZ6R/w18ZlXSHBYXiYUPO3g=="],
 
     "readable-stream/buffer": ["buffer@6.0.3", "", { "dependencies": { "base64-js": "^1.3.1", "ieee754": "^1.2.1" } }, "sha512-FTiCpNxtwiZZHEZbcbTIcZjERVICn9yq/pDFkTl95/AxzD1naBctN7YO68riM/gLSDY7sdrMby8hofADYuuqOA=="],
@@ -6663,6 +6673,14 @@
 
     "@opencode-ai/web/@shikijs/transformers/@shikijs/types": ["@shikijs/types@3.20.0", "", { "dependencies": { "@shikijs/vscode-textmate": "^10.0.2", "@types/hast": "^3.0.4" } }, "sha512-lhYAATn10nkZcBQ0BlzSbJA3wcmL5MXUUF8d2Zzon6saZDlToKaiRX60n2+ZaHJCmXEcZRWNzn+k9vplr8Jhsw=="],
 
+    "@opentelemetry/otlp-transformer/protobufjs/@protobufjs/codegen": ["@protobufjs/codegen@2.0.4", "", {}, "sha512-YyFaikqM5sH0ziFZCN3xDC7zeGaB/d0IUb9CATugHWbd1FRFwWwt4ld4OYMPWu5a3Xe01mGAULCdqhMlPl29Jg=="],
+
+    "@opentelemetry/otlp-transformer/protobufjs/@protobufjs/inquire": ["@protobufjs/inquire@1.1.0", "", {}, "sha512-kdSefcPdruJiFMVSbn801t4vFK7KB/5gd2fYvrxhuJYg8ILrmn9SKSX2tZdV6V+ksulWqS7aXjBcRXl3wHoD9Q=="],
+
+    "@opentelemetry/otlp-transformer/protobufjs/@protobufjs/utf8": ["@protobufjs/utf8@1.1.0", "", {}, "sha512-Vvn3zZrhQZkkBE8LSuW3em98c0FwgO4nxzv6OdSxPKJIEKY2bGbHn+mhGIPerzI4twdxaP8/0+06HBpwf345Lw=="],
+
+    "@opentelemetry/otlp-transformer/protobufjs/@types/node": ["@types/node@22.13.9", "", { "dependencies": { "undici-types": "~6.20.0" } }, "sha512-acBjXdRJ3A6Pb3tqnw9HZmyR3Fiol3aGxRCK1x3d+6CDAMjl7I649wpSd+yNURCjbOUGu9tqtLKnTGxmK6CyGw=="],
+
     "@opentui/solid/@babel/core/semver": ["semver@6.3.1", "", { "bin": { "semver": "bin/semver.js" } }, "sha512-BR7VvDCVHO+q2xBEWskxS6DJE1qRnb7DxzUrogb71CWoSficBxYsiAGd+Kl0mmq/MprG9yArRkyrQxTO6XjMzA=="],
 
     "@pierre/diffs/@shikijs/transformers/@shikijs/core": ["@shikijs/core@3.20.0", "", { "dependencies": { "@shikijs/types": "3.20.0", "@shikijs/vscode-textmate": "^10.0.2", "@types/hast": "^3.0.4", "hast-util-to-html": "^9.0.5" } }, "sha512-f2ED7HYV4JEk827mtMDwe/yQ25pRiXZmtHjWF8uzZKuKiEsJR7Ce1nuQ+HhV9FzDcbIo4ObBCD9GPTzNuy9S1g=="],
@@ -6959,8 +6977,6 @@
 
     "pkg-up/find-up/locate-path": ["locate-path@3.0.0", "", { "dependencies": { "p-locate": "^3.0.0", "path-exists": "^3.0.0" } }, "sha512-7AO748wWnIhNqAuaty2ZWHkQHRSNfPVIsPIfwEOWO22AmaoVrWavlOcMR5nzTLNYvp36X220/maaRsrec1G65A=="],
 
-    "protobufjs/@types/node/undici-types": ["undici-types@6.20.0", "", {}, "sha512-Ny6QZ2Nju20vw1SRHe3d9jVu6gJ+4e3+MMpqu7pqE5HT6WsTSlce++GQmK5UXS8mzV8DSYHrQH+Xrf2jVcuKNg=="],
-
     "readable-stream/buffer/ieee754": ["ieee754@1.2.1", "", {}, "sha512-dcyqhDvX1C46lXZcVqCpK+FtMRQVdIMN6/Df5js2zouUsqG7I6sFxitIC+7KYK29KdXOLHdu9zL4sFnoVQnqaA=="],
 
     "readdir-glob/minimatch/brace-expansion": ["brace-expansion@2.1.0", "", { "dependencies": { "balanced-match": "^1.0.0" } }, "sha512-TN1kCZAgdgweJhWWpgKYrQaMNHcDULHkWwQIspdtjV4Y5aurRdZpjAqn6yX3FPqTA9ngHCc4hJxMAMgGfve85w=="],
@@ -7213,6 +7229,8 @@
 
     "@opencode-ai/console-function/@ai-sdk/openai/@ai-sdk/provider-utils/@standard-schema/spec": ["@standard-schema/spec@1.1.0", "", {}, "sha512-l2aFy5jALhniG5HgqrD6jXLi/rUWrKvqN/qJx6yoJsgKhblVd+iqqU4RCXavm/jPityDo5TCvKMnpjKnOriy0w=="],
 
+    "@opentelemetry/otlp-transformer/protobufjs/@types/node/undici-types": ["undici-types@6.20.0", "", {}, "sha512-Ny6QZ2Nju20vw1SRHe3d9jVu6gJ+4e3+MMpqu7pqE5HT6WsTSlce++GQmK5UXS8mzV8DSYHrQH+Xrf2jVcuKNg=="],
+
     "@sentry/bundler-plugin-core/glob/minimatch/brace-expansion": ["brace-expansion@2.1.0", "", { "dependencies": { "balanced-match": "^1.0.0" } }, "sha512-TN1kCZAgdgweJhWWpgKYrQaMNHcDULHkWwQIspdtjV4Y5aurRdZpjAqn6yX3FPqTA9ngHCc4hJxMAMgGfve85w=="],
 
     "@sentry/bundler-plugin-core/glob/path-scurry/lru-cache": ["lru-cache@10.4.3", "", {}, "sha512-JNAzZcXrCt42VGLuYz0zfAzDfAvJWW6AfYlDBQyDV5DClI2m5sAmK+OIO7s59XfsRsWHp02jAJrRadPRGTt6SQ=="],
diff --git a/infra/app.ts b/infra/app.ts
index bb627f51ec..2ede5a1f4a 100644
--- a/infra/app.ts
+++ b/infra/app.ts
@@ -30,6 +30,7 @@ export const api = new sst.cloudflare.Worker("Api", {
   transform: {
     worker: (args) => {
       args.logpush = true
+      if ($app.stage === "vimtor") return
       args.bindings = $resolve(args.bindings).apply((bindings) => [
         ...bindings,
         {
diff --git a/infra/console.ts b/infra/console.ts
index 201d5bdc65..ab6502a8f8 100644
--- a/infra/console.ts
+++ b/infra/console.ts
@@ -1,5 +1,6 @@
 import { domain } from "./stage"
 import { EMAILOCTOPUS_API_KEY } from "./app"
+import { SECRET } from "./secret"
 
 ////////////////
 // DATABASE
@@ -221,6 +222,7 @@ const AUTH_API_URL = new sst.Linkable("AUTH_API_URL", {
 const STRIPE_WEBHOOK_SECRET = new sst.Linkable("STRIPE_WEBHOOK_SECRET", {
   properties: { value: stripeWebhook.secret },
 })
+
 const gatewayKv = new sst.cloudflare.Kv("GatewayKv")
 
 ////////////////
@@ -230,6 +232,7 @@ const gatewayKv = new sst.cloudflare.Kv("GatewayKv")
 const bucket = new sst.cloudflare.Bucket("ZenData")
 const bucketNew = new sst.cloudflare.Bucket("ZenDataNew")
 
+const DISCORD_INCIDENT_WEBHOOK_URL = new sst.Secret("DISCORD_INCIDENT_WEBHOOK_URL")
 const AWS_SES_ACCESS_KEY_ID = new sst.Secret("AWS_SES_ACCESS_KEY_ID")
 const AWS_SES_SECRET_ACCESS_KEY = new sst.Secret("AWS_SES_SECRET_ACCESS_KEY")
 
@@ -251,6 +254,8 @@ new sst.cloudflare.x.SolidStart("Console", {
     database,
     AUTH_API_URL,
     STRIPE_WEBHOOK_SECRET,
+    DISCORD_INCIDENT_WEBHOOK_URL,
+    SECRET.HoneycombWebhookSecret,
     STRIPE_SECRET_KEY,
     EMAILOCTOPUS_API_KEY,
     AWS_SES_ACCESS_KEY_ID,
diff --git a/infra/monitoring.ts b/infra/monitoring.ts
new file mode 100644
index 0000000000..aad090aa80
--- /dev/null
+++ b/infra/monitoring.ts
@@ -0,0 +1,199 @@
+import { SECRET } from "./secret"
+import { domain } from "./stage"
+
+const webhookRecipient = new honeycomb.WebhookRecipient("DiscordAlerts", {
+  name: $app.stage === "production" ? "Discord Alerts" : `Discord Alerts (${$app.stage})`,
+  url: `https://${domain}/honeycomb/webhook`,
+  secret: SECRET.HoneycombWebhookSecret.result,
+  templates: [
+    {
+      type: "trigger",
+      body: `{
+        "url": {{ .Result.URL | quote }},
+        "type": {{ .Vars.type | quote }},
+        "name": {{ .Name | quote }},
+        "status": {{ .Alert.Status | quote }},
+        "isTest": {{ .Alert.IsTest }},
+        "groups": {{ .Result.GroupsTriggered | toJson }}
+      }`,
+    },
+  ],
+  variables: [
+    {
+      name: "type",
+    },
+  ],
+})
+
+const modelHttpErrorsQuery = (product: "go" | "zen") => {
+  const filters = [
+    { column: "model", op: "exists" },
+    { column: "event_type", op: "=", value: "completions" },
+    { column: "user_agent", op: "contains", value: "opencode" },
+    { column: "isGoTier", op: "=", value: product === "go" ? "true" : "false" },
+  ]
+
+  return honeycomb.getQuerySpecificationOutput({
+    breakdowns: ["model"],
+    calculatedFields: [
+      {
+        name: "is_failed_http_status",
+        expression: `IF(AND(GTE($status, "400"), NOT(EQUALS($status, "401"))), 1, 0)`,
+      },
+    ],
+    calculations: [
+      { op: "COUNT", name: "TOTAL", filterCombination: "AND", filters },
+      { op: "SUM", name: "FAILED", column: "is_failed_http_status", filterCombination: "AND", filters },
+    ],
+    formulas: [{ name: "ERROR", expression: "IF(GTE($TOTAL, 500), DIV($FAILED, $TOTAL), 0)" }],
+    timeRange: 900,
+  }).json
+}
+
+const providerHttpErrorsQuery = (product: "go" | "zen") => {
+  const filters = [
+    { column: "provider", op: "exists" },
+    { column: "user_agent", op: "contains", value: "opencode" },
+    { column: "isGoTier", op: "=", value: product === "go" ? "true" : "false" },
+  ]
+
+  return honeycomb.getQuerySpecificationOutput({
+    breakdowns: ["provider"],
+    calculatedFields: [
+      {
+        name: "is_success_http_status",
+        expression: `IF(AND(GTE($status, "200"), LT($status, "400")), 1, 0)`,
+      },
+      {
+        name: "is_failed_provider_http_status",
+        expression: `IF(GTE($llm.error.code, "400"), 1, 0)`,
+      },
+    ],
+    calculations: [
+      {
+        op: "SUM",
+        name: "SUCCESS",
+        column: "is_success_http_status",
+        filterCombination: "AND",
+        filters: [...filters, { column: "event_type", op: "=", value: "completions" }],
+      },
+      {
+        op: "SUM",
+        name: "FAILED",
+        column: "is_failed_provider_http_status",
+        filterCombination: "AND",
+        filters: [...filters, { column: "event_type", op: "=", value: "llm.error" }],
+      },
+    ],
+    formulas: [
+      { name: "ERROR", expression: "IF(GTE(SUM($SUCCESS, $FAILED), 250), DIV($FAILED, SUM($SUCCESS, $FAILED)), 0)" },
+    ],
+    timeRange: 1800,
+  }).json
+}
+
+const description = "Managed by SST (Don't edit in Honeycomb UI)"
+
+new honeycomb.Trigger("IncreasedModelHttpErrorsGo", {
+  name: "Increased Model HTTP Errors [Go]",
+  description,
+  queryJson: modelHttpErrorsQuery("go"),
+  alertType: "on_change",
+  frequency: 300,
+  thresholds: [{ op: ">=", value: 0.8, exceededLimit: 1 }],
+  recipients: [
+    {
+      id: webhookRecipient.id,
+      notificationDetails: [
+        {
+          variables: [{ name: "type", value: "model_http_errors" }],
+        },
+      ],
+    },
+  ],
+})
+
+new honeycomb.Trigger("IncreasedModelHttpErrorsZen", {
+  name: "Increased Model HTTP Errors [Zen]",
+  description,
+  queryJson: modelHttpErrorsQuery("zen"),
+  alertType: "on_change",
+  frequency: 300,
+  thresholds: [{ op: ">=", value: 0.8, exceededLimit: 1 }],
+  recipients: [
+    {
+      id: webhookRecipient.id,
+      notificationDetails: [
+        {
+          variables: [{ name: "type", value: "model_http_errors" }],
+        },
+      ],
+    },
+  ],
+})
+
+new honeycomb.Trigger("IncreasedProviderHttpErrorsGo", {
+  name: "Increased Provider HTTP Errors [Go]",
+  description,
+  queryJson: providerHttpErrorsQuery("go"),
+  alertType: "on_change",
+  frequency: 600,
+  thresholds: [{ op: ">=", value: 0.8, exceededLimit: 1 }],
+  recipients: [
+    {
+      id: webhookRecipient.id,
+      notificationDetails: [
+        {
+          variables: [{ name: "type", value: "provider_http_errors" }],
+        },
+      ],
+    },
+  ],
+})
+
+new honeycomb.Trigger("IncreasedProviderHttpErrorsZen", {
+  name: "Increased Provider HTTP Errors [Zen]",
+  description,
+  queryJson: providerHttpErrorsQuery("zen"),
+  alertType: "on_change",
+  frequency: 600,
+  thresholds: [{ op: ">=", value: 0.8, exceededLimit: 1 }],
+  recipients: [
+    {
+      id: webhookRecipient.id,
+      notificationDetails: [
+        {
+          variables: [{ name: "type", value: "provider_http_errors" }],
+        },
+      ],
+    },
+  ],
+})
+
+new honeycomb.Trigger("IncreasedFreeTierRequests", {
+  name: "Increased Free Tier Requests",
+  description,
+  queryJson: honeycomb.getQuerySpecificationOutput({
+    calculations: [{ op: "COUNT" }],
+    filters: [
+      { column: "event_type", op: "=", value: "completions" },
+      { column: "user_agent", op: "contains", value: "opencode" },
+      { column: "isFreeTier", op: "=", value: "true" },
+    ],
+    timeRange: 3600,
+  }).json,
+  alertType: "on_change",
+  frequency: 900,
+  thresholds: [{ op: ">=", value: 60, exceededLimit: 1 }],
+  baselineDetails: [{ type: "percentage", offsetMinutes: 1440 }],
+  recipients: [
+    {
+      id: webhookRecipient.id,
+      notificationDetails: [
+        {
+          variables: [{ name: "type", value: "custom" }],
+        },
+      ],
+    },
+  ],
+})
diff --git a/infra/secret.ts b/infra/secret.ts
index 0b1870fa15..d4e8b148fc 100644
--- a/infra/secret.ts
+++ b/infra/secret.ts
@@ -1,4 +1,11 @@
+sst.Linkable.wrap(random.RandomPassword, (resource) => ({
+  properties: {
+    value: resource.result,
+  },
+}))
+
 export const SECRET = {
   R2AccessKey: new sst.Secret("R2AccessKey", "unknown"),
   R2SecretKey: new sst.Secret("R2SecretKey", "unknown"),
+  HoneycombWebhookSecret: new random.RandomPassword("HoneycombWebhookSecret", { length: 24 }),
 }
diff --git a/nix/hashes.json b/nix/hashes.json
index 441d0de8d9..078b600d05 100644
--- a/nix/hashes.json
+++ b/nix/hashes.json
@@ -1,8 +1,8 @@
 {
   "nodeModules": {
-    "x86_64-linux": "sha256-YBTnGuKDthi9wM4UrY0CMNqAzwnM6rN5XROyJOqYbQ8=",
-    "aarch64-linux": "sha256-7B1dxtYOc9t+e3lzF8O02YtjsvogyuZjHSanWw1XPio=",
-    "aarch64-darwin": "sha256-y0CzJRL4WHMxVbZPg3O7Dd+66TbITJbiv0oqhZ6URWw=",
-    "x86_64-darwin": "sha256-KW0Cx/ddKM4sQcpKhKwYu8qL6zYlm12kcUlgp66Wf50="
+    "x86_64-linux": "sha256-MHeO1KTmjYa+V4ZBYrQq93cYpjnkGfO9e3MOWwkzjVY=",
+    "aarch64-linux": "sha256-EqTRG7DrdKKT7CEvnaNk5VhjTRhlZ9juP9/Nnr3dJ+g=",
+    "aarch64-darwin": "sha256-c8dWd8Pgp5uIAOdYbHIeGKqWfkF/l4Ze7ArYUMvTNkE=",
+    "x86_64-darwin": "sha256-61NpSO0AZ4iZG19RQ6zg0SJec+VQE46WJKOdRrNofT0="
   }
 }
diff --git a/packages/app/package.json b/packages/app/package.json
index def3f65fc2..600c011b6b 100644
--- a/packages/app/package.json
+++ b/packages/app/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@opencode-ai/app",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "description": "",
   "type": "module",
   "exports": {
diff --git a/packages/app/src/components/dialog-select-mcp.tsx b/packages/app/src/components/dialog-select-mcp.tsx
index 9bb36d32d8..576ec8fec4 100644
--- a/packages/app/src/components/dialog-select-mcp.tsx
+++ b/packages/app/src/components/dialog-select-mcp.tsx
@@ -6,7 +6,7 @@ import { Dialog } from "@opencode-ai/ui/dialog"
 import { List } from "@opencode-ai/ui/list"
 import { Switch } from "@opencode-ai/ui/switch"
 import { useLanguage } from "@/context/language"
-import { loadMcpQuery } from "@/context/global-sync"
+import { mcpQueryKey } from "@/context/global-sync"
 
 const statusLabels = {
   connected: "mcp.status.connected",
@@ -32,7 +32,7 @@ export const DialogSelectMcp: Component = () => {
       if (sync.data.mcp[name]?.status === "connected") await sdk.client.mcp.disconnect({ name })
       else await sdk.client.mcp.connect({ name })
     },
-    onSuccess: () => queryClient.refetchQueries({ queryKey: loadMcpQuery(sync.directory).queryKey }),
+    onSuccess: () => queryClient.refetchQueries({ queryKey: mcpQueryKey(sync.directory) }),
   }))
 
   const enabledCount = createMemo(() => items().filter((i) => i.status === "connected").length)
diff --git a/packages/app/src/components/prompt-input.tsx b/packages/app/src/components/prompt-input.tsx
index 0a18096164..2417fa98e2 100644
--- a/packages/app/src/components/prompt-input.tsx
+++ b/packages/app/src/components/prompt-input.tsx
@@ -16,6 +16,7 @@ import {
 } from "@/context/prompt"
 import { useLayout } from "@/context/layout"
 import { useSDK } from "@/context/sdk"
+import { useGlobalSDK } from "@/context/global-sdk"
 import { useSync } from "@/context/sync"
 import { useComments } from "@/context/comments"
 import { Button } from "@opencode-ai/ui/button"
@@ -102,6 +103,7 @@ const NON_EMPTY_TEXT = /[^\s\u200B]/
 
 export const PromptInput: Component<PromptInputProps> = (props) => {
   const sdk = useSDK()
+  const globalSDK = useGlobalSDK()
 
   const sync = useSync()
   const local = useLocal()
@@ -1253,7 +1255,11 @@ export const PromptInput: Component<PromptInputProps> = (props) => {
   }
 
   const [agentsQuery, globalProvidersQuery, providersQuery] = useQueries(() => ({
-    queries: [loadAgentsQuery(sdk.directory), loadProvidersQuery(null), loadProvidersQuery(sdk.directory)],
+    queries: [
+      loadAgentsQuery(sdk.directory, sdk.client),
+      loadProvidersQuery(null, globalSDK.client),
+      loadProvidersQuery(sdk.directory, sdk.client),
+    ],
   }))
 
   const agentsLoading = () => agentsQuery.isLoading
diff --git a/packages/app/src/components/status-popover-body.tsx b/packages/app/src/components/status-popover-body.tsx
index 952e3eac64..bbac562784 100644
--- a/packages/app/src/components/status-popover-body.tsx
+++ b/packages/app/src/components/status-popover-body.tsx
@@ -15,7 +15,7 @@ import { useSDK } from "@/context/sdk"
 import { normalizeServerUrl, ServerConnection, useServer } from "@/context/server"
 import { useSync } from "@/context/sync"
 import { useCheckServerHealth, type ServerHealth } from "@/utils/server-health"
-import { loadMcpQuery } from "@/context/global-sync"
+import { mcpQueryKey } from "@/context/global-sync"
 
 const pollMs = 10_000
 
@@ -145,7 +145,7 @@ const useMcpToggleMutation = () => {
       const status = sync.data.mcp[name]
       await (status?.status === "connected" ? sdk.client.mcp.disconnect({ name }) : sdk.client.mcp.connect({ name }))
     },
-    onSuccess: () => queryClient.refetchQueries({ queryKey: loadMcpQuery(sync.directory).queryKey }),
+    onSuccess: () => queryClient.refetchQueries({ queryKey: mcpQueryKey(sync.directory) }),
     onError: (err) => {
       showToast({
         variant: "error",
diff --git a/packages/app/src/context/global-sync.tsx b/packages/app/src/context/global-sync.tsx
index 6190deb1ee..31c90463d8 100644
--- a/packages/app/src/context/global-sync.tsx
+++ b/packages/app/src/context/global-sync.tsx
@@ -20,7 +20,6 @@ import {
   clearProviderRev,
   loadGlobalConfigQuery,
   loadPathQuery,
-  loadProjectsQuery,
   loadProvidersQuery,
 } from "./global-sync/bootstrap"
 import { createChildStoreManager } from "./global-sync/child-store"
@@ -31,7 +30,7 @@ import { trimSessions } from "./global-sync/session-trim"
 import type { ProjectMeta } from "./global-sync/types"
 import { SESSION_RECENT_LIMIT } from "./global-sync/types"
 import { formatServerError } from "@/utils/server-errors"
-import { queryOptions, skipToken, useMutation, useQueries, useQuery, useQueryClient } from "@tanstack/solid-query"
+import { queryOptions, useMutation, useQueries, useQuery, useQueryClient } from "@tanstack/solid-query"
 import { createRefreshQueue } from "./global-sync/queue"
 import { directoryKey } from "./global-sync/utils"
 
@@ -49,19 +48,22 @@ type GlobalStore = {
   reload: undefined | "pending" | "complete"
 }
 
-export const loadSessionsQuery = (directory: string) =>
-  queryOptions<null>({ queryKey: [directory, "loadSessions"], queryFn: skipToken })
+export const loadSessionsQueryKey = (directory: string) => [directory, "loadSessions"] as const
 
-export const loadMcpQuery = (directory: string, sdk?: OpencodeClient) =>
+export const mcpQueryKey = (directory: string) => [directory, "mcp"] as const
+
+export const loadMcpQuery = (directory: string, sdk: OpencodeClient) =>
   queryOptions({
-    queryKey: [directory, "mcp"],
-    queryFn: sdk ? () => sdk.mcp.status().then((r) => r.data ?? {}) : skipToken,
+    queryKey: mcpQueryKey(directory),
+    queryFn: () => sdk.mcp.status().then((r) => r.data ?? {}),
   })
 
-export const loadLspQuery = (directory: string, sdk?: OpencodeClient) =>
+export const lspQueryKey = (directory: string) => [directory, "lsp"] as const
+
+export const loadLspQuery = (directory: string, sdk: OpencodeClient) =>
   queryOptions({
-    queryKey: [directory, "lsp"],
-    queryFn: sdk ? () => sdk.lsp.status().then((r) => r.data ?? []) : skipToken,
+    queryKey: lspQueryKey(directory),
+    queryFn: () => sdk.lsp.status().then((r) => r.data ?? []),
   })
 
 function createGlobalSync() {
@@ -76,7 +78,11 @@ function createGlobalSync() {
   const sessionMeta = new Map<string, { limit: number }>()
 
   const [configQuery, providerQuery, pathQuery] = useQueries(() => ({
-    queries: [loadGlobalConfigQuery(), loadProvidersQuery(null), loadPathQuery(null), loadProjectsQuery()],
+    queries: [
+      loadGlobalConfigQuery(globalSDK.client),
+      loadProvidersQuery(null, globalSDK.client),
+      loadPathQuery(null, globalSDK.client),
+    ],
   }))
 
   const [globalStore, setGlobalStore] = createStore<GlobalStore>({
@@ -233,7 +239,7 @@ function createGlobalSync() {
     const limit = Math.max(store.limit + SESSION_RECENT_LIMIT, SESSION_RECENT_LIMIT)
     const promise = queryClient
       .fetchQuery({
-        ...loadSessionsQuery(key),
+        queryKey: loadSessionsQueryKey(key),
         queryFn: () =>
           loadRootSessionsWithFallback({
             directory,
diff --git a/packages/app/src/context/global-sync/bootstrap.ts b/packages/app/src/context/global-sync/bootstrap.ts
index e85516bf14..531917bde6 100644
--- a/packages/app/src/context/global-sync/bootstrap.ts
+++ b/packages/app/src/context/global-sync/bootstrap.ts
@@ -18,7 +18,7 @@ import { reconcile, type SetStoreFunction, type Store } from "solid-js/store"
 import type { State, VcsCache } from "./types"
 import { cmp, normalizeAgentList, normalizeProviderList } from "./utils"
 import { formatServerError } from "@/utils/server-errors"
-import { QueryClient, queryOptions, skipToken } from "@tanstack/solid-query"
+import { QueryClient, queryOptions } from "@tanstack/solid-query"
 import { loadMcpQuery } from "../global-sync"
 
 type GlobalStore = {
@@ -83,44 +83,25 @@ function showErrors(input: {
   })
 }
 
-export const loadGlobalConfigQuery = (
-  sdk?: OpencodeClient,
-  transform?: (x: Awaited<ReturnType<OpencodeClient["global"]["config"]["get"]>>) => void,
-) =>
+export const loadGlobalConfigQuery = (sdk: OpencodeClient) =>
   queryOptions({
     queryKey: ["config"],
-    queryFn: sdk
-      ? () =>
-          retry(() =>
-            sdk.global.config.get().then((x) => {
-              transform?.(x)
-              return x.data!
-            }),
-          )
-      : skipToken,
+    queryFn: () => retry(() => sdk.global.config.get().then((x) => x.data!)),
   })
 
-export const loadProjectsQuery = (
-  sdk?: OpencodeClient,
-  transform?: (x: Awaited<ReturnType<OpencodeClient["project"]["list"]>>["data"]) => void,
-) =>
+export const loadProjectsQuery = (sdk: OpencodeClient) =>
   queryOptions({
     queryKey: ["project"],
-    queryFn: sdk
-      ? () =>
-          retry(() =>
-            sdk.project
-              .list()
-              .then((x) => {
-                return (x.data ?? [])
-                  .filter((p) => !!p?.id)
-                  .filter((p) => !!p.worktree && !p.worktree.includes("opencode-test"))
-                  .slice()
-                  .sort((a, b) => cmp(a.id, b.id))
-              })
-              .then(transform),
-          )
-      : skipToken,
+    queryFn: () =>
+      retry(() =>
+        sdk.project.list().then((x) => {
+          return (x.data ?? [])
+            .filter((p) => !!p?.id)
+            .filter((p) => !!p.worktree && !p.worktree.includes("opencode-test"))
+            .slice()
+            .sort((a, b) => cmp(a.id, b.id))
+        }),
+      ),
   })
 
 export async function bootstrapGlobal(input: {
@@ -136,9 +117,9 @@ export async function bootstrapGlobal(input: {
     () => input.queryClient.fetchQuery(loadProvidersQuery(null, input.globalSDK)),
     () => input.queryClient.fetchQuery(loadPathQuery(null, input.globalSDK)),
     () =>
-      input.queryClient.fetchQuery(
-        loadProjectsQuery(input.globalSDK, (data) => input.setGlobalStore("project", data ?? [])),
-      ),
+      input.queryClient
+        .fetchQuery(loadProjectsQuery(input.globalSDK))
+        .then((data) => input.setGlobalStore("project", data)),
   ]
   await runAll(slow)
   // showErrors({
@@ -197,46 +178,22 @@ function warmSessions(input: {
   ).then(() => undefined)
 }
 
-export const loadProvidersQuery = (directory: string | null, sdk?: OpencodeClient) =>
+export const loadProvidersQuery = (directory: string | null, sdk: OpencodeClient) =>
   queryOptions({
     queryKey: [directory, "providers"],
-    queryFn: sdk ? () => retry(() => sdk.provider.list().then((x) => normalizeProviderList(x.data!))) : skipToken,
+    queryFn: () => retry(() => sdk.provider.list().then((x) => normalizeProviderList(x.data!))),
   })
 
-export const loadAgentsQuery = (
-  directory: string | null,
-  sdk?: OpencodeClient,
-  transform?: (x: Awaited<ReturnType<OpencodeClient["app"]["agents"]>>) => void,
-) =>
+export const loadAgentsQuery = (directory: string | null, sdk: OpencodeClient) =>
   queryOptions({
     queryKey: [directory, "agents"],
-    queryFn: sdk
-      ? () =>
-          retry(() =>
-            sdk.app.agents().then((x) => {
-              transform?.(x)
-              return x.data!
-            }),
-          )
-      : skipToken,
+    queryFn: () => retry(() => sdk.app.agents().then((x) => normalizeAgentList(x.data))),
   })
 
-export const loadPathQuery = (
-  directory: string | null,
-  sdk?: OpencodeClient,
-  transform?: (x: Awaited<ReturnType<OpencodeClient["path"]["get"]>>) => void,
-) =>
+export const loadPathQuery = (directory: string | null, sdk: OpencodeClient) =>
   queryOptions<Path>({
     queryKey: [directory, "path"],
-    queryFn: sdk
-      ? () =>
-          retry(() =>
-            sdk.path.get().then(async (x) => {
-              transform?.(x)
-              return x.data!
-            }),
-          )
-      : skipToken,
+    queryFn: () => retry(() => sdk.path.get().then((x) => x.data!)),
   })
 
 export async function bootstrapDirectory(input: {
@@ -271,9 +228,9 @@ export async function bootstrapDirectory(input: {
     const slow = [
       () => Promise.resolve(input.loadSessions(input.directory)),
       () =>
-        input.queryClient.ensureQueryData(
-          loadAgentsQuery(input.directory, input.sdk, (x) => input.setStore("agent", normalizeAgentList(x.data))),
-        ),
+        input.queryClient
+          .ensureQueryData(loadAgentsQuery(input.directory, input.sdk))
+          .then((data) => input.setStore("agent", data)),
       () =>
         retry(() => input.sdk.config.get().then((x) => input.setStore("config", reconcile(x.data!, { merge: false })))),
       () => retry(() => input.sdk.session.status().then((x) => input.setStore("session_status", x.data!))),
@@ -281,12 +238,10 @@ export async function bootstrapDirectory(input: {
         (() => retry(() => input.sdk.project.current()).then((x) => input.setStore("project", x.data!.id))),
       !seededPath &&
         (() =>
-          input.queryClient.ensureQueryData(
-            loadPathQuery(input.directory, input.sdk, (x) => {
-              const next = projectID(x.data?.directory ?? input.directory, input.global.project)
-              if (next) input.setStore("project", next)
-            }),
-          )),
+          input.queryClient.ensureQueryData(loadPathQuery(input.directory, input.sdk)).then((data) => {
+            const next = projectID(data.directory ?? input.directory, input.global.project)
+            if (next) input.setStore("project", next)
+          })),
       () =>
         retry(() =>
           input.sdk.vcs.get().then((x) => {
diff --git a/packages/app/src/pages/layout/sidebar-workspace.tsx b/packages/app/src/pages/layout/sidebar-workspace.tsx
index d2e887b444..9b80adac29 100644
--- a/packages/app/src/pages/layout/sidebar-workspace.tsx
+++ b/packages/app/src/pages/layout/sidebar-workspace.tsx
@@ -14,12 +14,12 @@ import { Spinner } from "@opencode-ai/ui/spinner"
 import { Tooltip } from "@opencode-ai/ui/tooltip"
 import { type Session } from "@opencode-ai/sdk/v2/client"
 import { type LocalProject } from "@/context/layout"
-import { loadSessionsQuery, useGlobalSync } from "@/context/global-sync"
+import { loadSessionsQueryKey, useGlobalSync } from "@/context/global-sync"
 import { useLanguage } from "@/context/language"
 import { pathKey } from "@/utils/path-key"
 import { NewSessionItem, SessionItem, SessionSkeleton } from "./sidebar-items"
 import { sortedRootSessions } from "./helpers"
-import { useQuery } from "@tanstack/solid-query"
+import { useIsFetching } from "@tanstack/solid-query"
 
 type InlineEditorComponent = (props: {
   id: string
@@ -320,9 +320,9 @@ export const SortableWorkspace = (props: {
   const boot = createMemo(() => open() || active())
   const count = createMemo(() => sessions()?.length ?? 0)
   const hasMore = createMemo(() => workspaceStore.sessionTotal > count())
-  const query = useQuery(() => ({ ...loadSessionsQuery(props.project.worktree) }))
+  const fetching = useIsFetching(() => ({ queryKey: loadSessionsQueryKey(props.directory) }))
   const busy = createMemo(() => props.ctx.isBusy(props.directory))
-  const loading = () => query.isLoading && count() === 0
+  const loading = () => fetching() > 0 && count() === 0
   const touch = createMediaQuery("(hover: none)")
   const showNew = createMemo(() => !loading() && (touch() || count() === 0 || (active() && !params.id)))
   const loadMore = async () => {
@@ -427,7 +427,7 @@ export const SortableWorkspace = (props: {
             mobile={props.mobile}
             ctx={props.ctx}
             showNew={showNew}
-            loading={() => query.isLoading && count() === 0}
+            loading={loading}
             sessions={sessions}
             hasMore={hasMore}
             loadMore={loadMore}
@@ -454,9 +454,9 @@ export const LocalWorkspace = (props: {
   const slug = createMemo(() => base64Encode(props.project.worktree))
   const sessions = createMemo(() => sortedRootSessions(workspace().store, props.sortNow()))
   const count = createMemo(() => sessions()?.length ?? 0)
-  const query = useQuery(() => ({ ...loadSessionsQuery(props.project.worktree) }))
+  const fetching = useIsFetching(() => ({ queryKey: loadSessionsQueryKey(props.project.worktree) }))
   const hasMore = createMemo(() => workspace().store.sessionTotal > count())
-  const loading = () => query.isLoading && count() === 0
+  const loading = () => fetching() > 0 && count() === 0
   const loadMore = async () => {
     workspace().setStore("limit", (limit) => (limit ?? 0) + 5)
     await globalSync.project.loadSessions(props.project.worktree)
diff --git a/packages/console/app/package.json b/packages/console/app/package.json
index 3d07a87cfd..f2471d2926 100644
--- a/packages/console/app/package.json
+++ b/packages/console/app/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@opencode-ai/console-app",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "type": "module",
   "license": "MIT",
   "scripts": {
diff --git a/packages/console/app/src/i18n/ar.ts b/packages/console/app/src/i18n/ar.ts
index 5c0919e8e2..12ec7f1fbd 100644
--- a/packages/console/app/src/i18n/ar.ts
+++ b/packages/console/app/src/i18n/ar.ts
@@ -249,7 +249,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | نماذج برمجة منخفضة التكلفة للجميع",
   "go.meta.description":
-    "يبدأ Go من $5 للشهر الأول، ثم $10/شهر، مع حدود طلب سخية لمدة 5 ساعات لـ GLM-5.1 وGLM-5 وKimi K2.5 وKimi K2.6 وMiMo-V2-Pro وMiMo-V2-Omni وMiMo-V2.5-Pro وMiMo-V2.5 وQwen3.5 Plus وQwen3.6 Plus وMiniMax M2.5 وMiniMax M2.7 وDeepSeek V4 Pro وDeepSeek V4 Flash.",
+    "يبدأ Go من $5 للشهر الأول، ثم $10/شهر، مع حدود طلب سخية لمدة 5 ساعات لـ GLM-5.1 وGLM-5 وKimi K2.5 وKimi K2.6 وMiMo-V2.5-Pro وMiMo-V2.5 وQwen3.5 Plus وQwen3.6 Plus وMiniMax M2.5 وMiniMax M2.7 وDeepSeek V4 Pro وDeepSeek V4 Flash.",
   "go.hero.title": "نماذج برمجة منخفضة التكلفة للجميع",
   "go.hero.body":
     "يجلب Go البرمجة الوكيلة للمبرمجين حول العالم. يوفر حدودًا سخية ووصولًا موثوقًا إلى أقوى النماذج مفتوحة المصدر، حتى تتمكن من البناء باستخدام وكلاء أقوياء دون القلق بشأن التكلفة أو التوفر.",
@@ -261,8 +261,6 @@ export const dict = {
   "go.cta.promo": "$5 للشهر الأول",
   "go.pricing.body":
     "استخدمه مع أي وكيل. $5 للشهر الأول، ثم $10/شهر. قم بزيادة الرصيد إذا لزم الأمر. الإلغاء في أي وقت.",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6: حد الاستخدام 3 أضعاف حتى 27 أبريل",
   "go.graph.free": "مجاني",
   "go.graph.freePill": "Big Pickle ونماذج مجانية",
   "go.graph.go": "Go",
@@ -300,7 +298,7 @@ export const dict = {
   "go.problem.item2": "حدود سخية ووصول موثوق",
   "go.problem.item3": "مصمم لأكبر عدد ممكن من المبرمجين",
   "go.problem.item4":
-    "يتضمن GLM-5.1 وGLM-5 وKimi K2.5 وKimi K2.6 وMiMo-V2-Pro وMiMo-V2-Omni وMiMo-V2.5-Pro وMiMo-V2.5 وQwen3.5 Plus وQwen3.6 Plus وMiniMax M2.5 وMiniMax M2.7 وDeepSeek V4 Pro وDeepSeek V4 Flash",
+    "يتضمن GLM-5.1 وGLM-5 وKimi K2.5 وKimi K2.6 وMiMo-V2.5-Pro وMiMo-V2.5 وQwen3.5 Plus وQwen3.6 Plus وMiniMax M2.5 وMiniMax M2.7 وDeepSeek V4 Pro وDeepSeek V4 Flash",
   "go.how.title": "كيف يعمل Go",
   "go.how.body": "يبدأ Go من $5 للشهر الأول، ثم $10/شهر. يمكنك استخدامه مع OpenCode أو أي وكيل.",
   "go.how.step1.title": "أنشئ حسابًا",
@@ -324,7 +322,7 @@ export const dict = {
   "go.faq.a2": "يتضمن Go النماذج المدرجة أدناه، مع حدود سخية وإتاحة موثوقة.",
   "go.faq.q3": "هل Go هو نفسه Zen؟",
   "go.faq.a3":
-    "لا. Zen هو الدفع حسب الاستخدام، بينما يبدأ Go من $5 للشهر الأول، ثم $10/شهر، مع حدود سخية ووصول موثوق إلى نماذج المصدر المفتوح GLM-5.1 وGLM-5 وKimi K2.5 وKimi K2.6 وMiMo-V2-Pro وMiMo-V2-Omni وMiMo-V2.5-Pro وMiMo-V2.5 وQwen3.5 Plus وQwen3.6 Plus وMiniMax M2.5 وMiniMax M2.7 وDeepSeek V4 Pro وDeepSeek V4 Flash.",
+    "لا. Zen هو الدفع حسب الاستخدام، بينما يبدأ Go من $5 للشهر الأول، ثم $10/شهر، مع حدود سخية ووصول موثوق إلى نماذج المصدر المفتوح GLM-5.1 وGLM-5 وKimi K2.5 وKimi K2.6 وMiMo-V2.5-Pro وMiMo-V2.5 وQwen3.5 Plus وQwen3.6 Plus وMiniMax M2.5 وMiniMax M2.7 وDeepSeek V4 Pro وDeepSeek V4 Flash.",
   "go.faq.q4": "كم تكلفة Go؟",
   "go.faq.a4.p1.beforePricing": "تكلفة Go",
   "go.faq.a4.p1.pricingLink": "$5 للشهر الأول",
@@ -347,7 +345,7 @@ export const dict = {
 
   "go.faq.q9": "ما الفرق بين النماذج المجانية وGo؟",
   "go.faq.a9":
-    "تشمل النماذج المجانية Big Pickle بالإضافة إلى النماذج الترويجية المتاحة في ذلك الوقت، مع حصة 200 طلب/يوم. يتضمن Go نماذج GLM-5.1 وGLM-5 وKimi K2.5 وKimi K2.6 وMiMo-V2-Pro وMiMo-V2-Omni وMiMo-V2.5-Pro وMiMo-V2.5 وQwen3.5 Plus وQwen3.6 Plus وMiniMax M2.5 وMiniMax M2.7 وDeepSeek V4 Pro وDeepSeek V4 Flash مع حصص طلبات أعلى مطبقة عبر نوافذ متجددة (5 ساعات، أسبوعيًا، وشهريًا)، تعادل تقريبًا 12 دولارًا كل 5 ساعات، و30 دولارًا في الأسبوع، و60 دولارًا في الشهر (تختلف أعداد الطلبات الفعلية حسب النموذج والاستخدام).",
+    "تشمل النماذج المجانية Big Pickle بالإضافة إلى النماذج الترويجية المتاحة في ذلك الوقت، مع حصة 200 طلب/يوم. يتضمن Go نماذج GLM-5.1 وGLM-5 وKimi K2.5 وKimi K2.6 وMiMo-V2.5-Pro وMiMo-V2.5 وQwen3.5 Plus وQwen3.6 Plus وMiniMax M2.5 وMiniMax M2.7 وDeepSeek V4 Pro وDeepSeek V4 Flash مع حصص طلبات أعلى مطبقة عبر نوافذ متجددة (5 ساعات، أسبوعيًا، وشهريًا)، تعادل تقريبًا 12 دولارًا كل 5 ساعات، و30 دولارًا في الأسبوع، و60 دولارًا في الشهر (تختلف أعداد الطلبات الفعلية حسب النموذج والاستخدام).",
 
   "zen.api.error.rateLimitExceeded": "تم تجاوز حد الطلبات. يرجى المحاولة مرة أخرى لاحقًا.",
   "zen.api.error.modelNotSupported": "النموذج {{model}} غير مدعوم",
diff --git a/packages/console/app/src/i18n/br.ts b/packages/console/app/src/i18n/br.ts
index 76e6987d3e..0a6d8f153e 100644
--- a/packages/console/app/src/i18n/br.ts
+++ b/packages/console/app/src/i18n/br.ts
@@ -253,7 +253,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | Modelos de codificação de baixo custo para todos",
   "go.meta.description":
-    "O Go começa em $5 no primeiro mês, depois $10/mês, com limites generosos de solicitação de 5 horas para GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash.",
+    "O Go começa em $5 no primeiro mês, depois $10/mês, com limites generosos de solicitação de 5 horas para GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash.",
   "go.hero.title": "Modelos de codificação de baixo custo para todos",
   "go.hero.body":
     "O Go traz a codificação com agentes para programadores em todo o mundo. Oferecendo limites generosos e acesso confiável aos modelos de código aberto mais capazes, para que você possa construir com agentes poderosos sem se preocupar com custos ou disponibilidade.",
@@ -265,8 +265,6 @@ export const dict = {
   "go.cta.promo": "$5 no primeiro mês",
   "go.pricing.body":
     "Use com qualquer agente. $5 no primeiro mês, depois $10/mês. Recarregue o crédito se necessário. Cancele a qualquer momento.",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6: limite de uso 3x maior até 27 de abril",
   "go.graph.free": "Grátis",
   "go.graph.freePill": "Big Pickle e modelos gratuitos",
   "go.graph.go": "Go",
@@ -305,7 +303,7 @@ export const dict = {
   "go.problem.item2": "Limites generosos e acesso confiável",
   "go.problem.item3": "Feito para o maior número possível de programadores",
   "go.problem.item4":
-    "Inclui GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash",
+    "Inclui GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash",
   "go.how.title": "Como o Go funciona",
   "go.how.body":
     "O Go começa em $5 no primeiro mês, depois $10/mês. Você pode usá-lo com o OpenCode ou qualquer agente.",
@@ -331,7 +329,7 @@ export const dict = {
   "go.faq.a2": "O Go inclui os modelos listados abaixo, com limites generosos e acesso confiável.",
   "go.faq.q3": "O Go é o mesmo que o Zen?",
   "go.faq.a3":
-    "Não. Zen é pay-as-you-go, enquanto o Go começa em $5 no primeiro mês, depois $10/mês, com limites generosos e acesso confiável aos modelos open source GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash.",
+    "Não. Zen é pay-as-you-go, enquanto o Go começa em $5 no primeiro mês, depois $10/mês, com limites generosos e acesso confiável aos modelos open source GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash.",
   "go.faq.q4": "Quanto custa o Go?",
   "go.faq.a4.p1.beforePricing": "O Go custa",
   "go.faq.a4.p1.pricingLink": "$5 no primeiro mês",
@@ -355,7 +353,7 @@ export const dict = {
 
   "go.faq.q9": "Qual a diferença entre os modelos gratuitos e o Go?",
   "go.faq.a9":
-    "Os modelos gratuitos incluem Big Pickle e modelos promocionais disponíveis no momento, com uma cota de 200 requisições/dia. O Go inclui GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash com cotas de requisição mais altas aplicadas em janelas móveis (5 horas, semanal e mensal), aproximadamente equivalentes a $12 por 5 horas, $30 por semana e $60 por mês (as contagens reais de requisições variam de acordo com o modelo e o uso).",
+    "Os modelos gratuitos incluem Big Pickle e modelos promocionais disponíveis no momento, com uma cota de 200 requisições/dia. O Go inclui GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash com cotas de requisição mais altas aplicadas em janelas móveis (5 horas, semanal e mensal), aproximadamente equivalentes a $12 por 5 horas, $30 por semana e $60 por mês (as contagens reais de requisições variam de acordo com o modelo e o uso).",
 
   "zen.api.error.rateLimitExceeded": "Limite de taxa excedido. Por favor, tente novamente mais tarde.",
   "zen.api.error.modelNotSupported": "Modelo {{model}} não suportado",
diff --git a/packages/console/app/src/i18n/da.ts b/packages/console/app/src/i18n/da.ts
index b97ee2cc0a..15e7151b67 100644
--- a/packages/console/app/src/i18n/da.ts
+++ b/packages/console/app/src/i18n/da.ts
@@ -251,7 +251,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | Kodningsmodeller til lav pris for alle",
   "go.meta.description":
-    "Go starter ved $5 for den første måned, derefter $10/måned, med generøse 5-timers anmodningsgrænser for GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash.",
+    "Go starter ved $5 for den første måned, derefter $10/måned, med generøse 5-timers anmodningsgrænser for GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash.",
   "go.hero.title": "Kodningsmodeller til lav pris for alle",
   "go.hero.body":
     "Go bringer agentisk kodning til programmører over hele verden. Med generøse grænser og pålidelig adgang til de mest kapable open source-modeller, så du kan bygge med kraftfulde agenter uden at bekymre dig om omkostninger eller tilgængelighed.",
@@ -263,8 +263,6 @@ export const dict = {
   "go.cta.promo": "$5 første måned",
   "go.pricing.body":
     "Brug med enhver agent. $5 første måned, derefter $10/måned. Tank op med kredit efter behov. Afmeld når som helst.",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6: brugsgrænsen tredoblet til 27. april",
   "go.graph.free": "Gratis",
   "go.graph.freePill": "Big Pickle og gratis modeller",
   "go.graph.go": "Go",
@@ -302,7 +300,7 @@ export const dict = {
   "go.problem.item2": "Generøse grænser og pålidelig adgang",
   "go.problem.item3": "Bygget til så mange programmører som muligt",
   "go.problem.item4":
-    "Inkluderer GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash",
+    "Inkluderer GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash",
   "go.how.title": "Hvordan Go virker",
   "go.how.body":
     "Go starter ved $5 for den første måned, derefter $10/måned. Du kan bruge det med OpenCode eller enhver agent.",
@@ -328,7 +326,7 @@ export const dict = {
   "go.faq.a2": "Go inkluderer modellerne nedenfor med generøse grænser og pålidelig adgang.",
   "go.faq.q3": "Er Go det samme som Zen?",
   "go.faq.a3":
-    "Nej. Zen er pay-as-you-go, mens Go starter ved $5 for den første måned, derefter $10/måned, med generøse grænser og pålidelig adgang til open source-modellerne GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash.",
+    "Nej. Zen er pay-as-you-go, mens Go starter ved $5 for den første måned, derefter $10/måned, med generøse grænser og pålidelig adgang til open source-modellerne GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash.",
   "go.faq.q4": "Hvad koster Go?",
   "go.faq.a4.p1.beforePricing": "Go koster",
   "go.faq.a4.p1.pricingLink": "$5 første måned",
@@ -351,7 +349,7 @@ export const dict = {
 
   "go.faq.q9": "Hvad er forskellen på gratis modeller og Go?",
   "go.faq.a9":
-    "Gratis modeller inkluderer Big Pickle plus salgsfremmende modeller tilgængelige på det tidspunkt, med en kvote på 200 forespørgsler/dag. Go inkluderer GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash med højere anmodningskvoter håndhævet over rullende vinduer (5-timers, ugentlig og månedlig), nogenlunde svarende til $12 pr. 5 timer, $30 pr. uge og $60 pr. måned (faktiske anmodningstal varierer efter model og brug).",
+    "Gratis modeller inkluderer Big Pickle plus salgsfremmende modeller tilgængelige på det tidspunkt, med en kvote på 200 forespørgsler/dag. Go inkluderer GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash med højere anmodningskvoter håndhævet over rullende vinduer (5-timers, ugentlig og månedlig), nogenlunde svarende til $12 pr. 5 timer, $30 pr. uge og $60 pr. måned (faktiske anmodningstal varierer efter model og brug).",
 
   "zen.api.error.rateLimitExceeded": "Hastighedsgrænse overskredet. Prøv venligst igen senere.",
   "zen.api.error.modelNotSupported": "Model {{model}} understøttes ikke",
diff --git a/packages/console/app/src/i18n/de.ts b/packages/console/app/src/i18n/de.ts
index 33b6e1b3de..0efcce78bf 100644
--- a/packages/console/app/src/i18n/de.ts
+++ b/packages/console/app/src/i18n/de.ts
@@ -253,7 +253,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | Kostengünstige Coding-Modelle für alle",
   "go.meta.description":
-    "Go beginnt bei $5 für den ersten Monat, danach $10/Monat, mit großzügigen 5-Stunden-Anfragelimits für GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro und DeepSeek V4 Flash.",
+    "Go beginnt bei $5 für den ersten Monat, danach $10/Monat, mit großzügigen 5-Stunden-Anfragelimits für GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro und DeepSeek V4 Flash.",
   "go.hero.title": "Kostengünstige Coding-Modelle für alle",
   "go.hero.body":
     "Go bringt Agentic Coding zu Programmierern auf der ganzen Welt. Mit großzügigen Limits und zuverlässigem Zugang zu den leistungsfähigsten Open-Source-Modellen, damit du mit leistungsstarken Agenten entwickeln kannst, ohne dir Gedanken über Kosten oder Verfügbarkeit zu machen.",
@@ -265,8 +265,6 @@ export const dict = {
   "go.cta.promo": "$5 im ersten Monat",
   "go.pricing.body":
     "Mit jedem Agenten nutzbar. $5 im ersten Monat, danach $10/Monat. Guthaben bei Bedarf aufladen. Jederzeit kündbar.",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6: Nutzungslimit bis zum 27. April verdreifacht",
   "go.graph.free": "Kostenlos",
   "go.graph.freePill": "Big Pickle und kostenlose Modelle",
   "go.graph.go": "Go",
@@ -304,7 +302,7 @@ export const dict = {
   "go.problem.item2": "Großzügige Limits und zuverlässiger Zugang",
   "go.problem.item3": "Für so viele Programmierer wie möglich gebaut",
   "go.problem.item4":
-    "Beinhaltet GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro und DeepSeek V4 Flash",
+    "Beinhaltet GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro und DeepSeek V4 Flash",
   "go.how.title": "Wie Go funktioniert",
   "go.how.body":
     "Go beginnt bei $5 für den ersten Monat, danach $10/Monat. Du kannst es mit OpenCode oder jedem Agenten nutzen.",
@@ -330,7 +328,7 @@ export const dict = {
   "go.faq.a2": "Go umfasst die unten aufgeführten Modelle mit großzügigen Limits und zuverlässigem Zugriff.",
   "go.faq.q3": "Ist Go dasselbe wie Zen?",
   "go.faq.a3":
-    "Nein. Zen ist Pay-as-you-go, während Go bei $5 für den ersten Monat beginnt, danach $10/Monat, mit großzügigen Limits und zuverlässigem Zugang zu den Open-Source-Modellen GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro und DeepSeek V4 Flash.",
+    "Nein. Zen ist Pay-as-you-go, während Go bei $5 für den ersten Monat beginnt, danach $10/Monat, mit großzügigen Limits und zuverlässigem Zugang zu den Open-Source-Modellen GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro und DeepSeek V4 Flash.",
   "go.faq.q4": "Wie viel kostet Go?",
   "go.faq.a4.p1.beforePricing": "Go kostet",
   "go.faq.a4.p1.pricingLink": "$5 im ersten Monat",
@@ -354,7 +352,7 @@ export const dict = {
 
   "go.faq.q9": "Was ist der Unterschied zwischen kostenlosen Modellen und Go?",
   "go.faq.a9":
-    "Kostenlose Modelle beinhalten Big Pickle sowie Werbemodelle, die zum jeweiligen Zeitpunkt verfügbar sind, mit einem Kontingent von 200 Anfragen/Tag. Go beinhaltet GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro und DeepSeek V4 Flash mit höheren Anfragekontingenten, die über rollierende Zeitfenster (5 Stunden, wöchentlich und monatlich) durchgesetzt werden, grob äquivalent zu $12 pro 5 Stunden, $30 pro Woche und $60 pro Monat (tatsächliche Anfragezahlen variieren je nach Modell und Nutzung).",
+    "Kostenlose Modelle beinhalten Big Pickle sowie Werbemodelle, die zum jeweiligen Zeitpunkt verfügbar sind, mit einem Kontingent von 200 Anfragen/Tag. Go beinhaltet GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro und DeepSeek V4 Flash mit höheren Anfragekontingenten, die über rollierende Zeitfenster (5 Stunden, wöchentlich und monatlich) durchgesetzt werden, grob äquivalent zu $12 pro 5 Stunden, $30 pro Woche und $60 pro Monat (tatsächliche Anfragezahlen variieren je nach Modell und Nutzung).",
 
   "zen.api.error.rateLimitExceeded": "Ratenlimit überschritten. Bitte versuche es später erneut.",
   "zen.api.error.modelNotSupported": "Modell {{model}} wird nicht unterstützt",
diff --git a/packages/console/app/src/i18n/en.ts b/packages/console/app/src/i18n/en.ts
index b6934b94de..f2cf3c14a4 100644
--- a/packages/console/app/src/i18n/en.ts
+++ b/packages/console/app/src/i18n/en.ts
@@ -248,9 +248,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | Low cost coding models for everyone",
   "go.meta.description":
-    "Go starts at $5 for your first month, then $10/month, with generous 5-hour request limits for GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, and DeepSeek V4 Flash.",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6 gets 3× usage limits through April 27",
+    "Go starts at $5 for your first month, then $10/month, with generous 5-hour request limits for GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, and DeepSeek V4 Flash.",
   "go.hero.title": "Low cost coding models for everyone",
   "go.hero.body":
     "Go brings agentic coding to programmers around the world. Offering generous limits and reliable access to the most capable open-source models, so you can build with powerful agents without worrying about cost or availability.",
@@ -298,7 +296,7 @@ export const dict = {
   "go.problem.item2": "Generous limits and reliable access",
   "go.problem.item3": "Built for as many programmers as possible",
   "go.problem.item4":
-    "Includes GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, and DeepSeek V4 Flash",
+    "Includes GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, and DeepSeek V4 Flash",
   "go.how.title": "How Go works",
   "go.how.body": "Go starts at $5 for your first month, then $10/month. You can use it with OpenCode or any agent.",
   "go.how.step1.title": "Create an account",
@@ -323,7 +321,7 @@ export const dict = {
   "go.faq.a2": "Go includes the models listed below, with generous limits and reliable access.",
   "go.faq.q3": "Is Go the same as Zen?",
   "go.faq.a3":
-    "No. Zen is pay-as-you-go, while Go starts at $5 for your first month, then $10/month, with generous limits and reliable access to open-source models GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, and DeepSeek V4 Flash.",
+    "No. Zen is pay-as-you-go, while Go starts at $5 for your first month, then $10/month, with generous limits and reliable access to open-source models GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, and DeepSeek V4 Flash.",
   "go.faq.q4": "How much does Go cost?",
   "go.faq.a4.p1.beforePricing": "Go costs",
   "go.faq.a4.p1.pricingLink": "$5 first month",
@@ -347,7 +345,7 @@ export const dict = {
 
   "go.faq.q9": "What is the difference between free models and Go?",
   "go.faq.a9":
-    "Free models include Big Pickle plus promotional models available at the time, with a quota of 200 requests/day. Go includes GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, and DeepSeek V4 Flash with higher request quotas enforced across rolling windows (5-hour, weekly, and monthly), roughly equivalent to $12 per 5 hours, $30 per week, and $60 per month (actual request counts vary by model and usage).",
+    "Free models include Big Pickle plus promotional models available at the time, with a quota of 200 requests/day. Go includes GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, and DeepSeek V4 Flash with higher request quotas enforced across rolling windows (5-hour, weekly, and monthly), roughly equivalent to $12 per 5 hours, $30 per week, and $60 per month (actual request counts vary by model and usage).",
 
   "zen.api.error.rateLimitExceeded": "Rate limit exceeded. Please try again later.",
   "zen.api.error.modelNotSupported": "Model {{model}} not supported",
diff --git a/packages/console/app/src/i18n/es.ts b/packages/console/app/src/i18n/es.ts
index c5cc71ae1e..5614a8c7ad 100644
--- a/packages/console/app/src/i18n/es.ts
+++ b/packages/console/app/src/i18n/es.ts
@@ -254,7 +254,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | Modelos de programación de bajo coste para todos",
   "go.meta.description":
-    "Go comienza en $5 el primer mes, luego 10 $/mes, con generosos límites de solicitudes de 5 horas para GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro y DeepSeek V4 Flash.",
+    "Go comienza en $5 el primer mes, luego 10 $/mes, con generosos límites de solicitudes de 5 horas para GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro y DeepSeek V4 Flash.",
   "go.hero.title": "Modelos de programación de bajo coste para todos",
   "go.hero.body":
     "Go lleva la programación agéntica a programadores de todo el mundo. Ofrece límites generosos y acceso fiable a los modelos de código abierto más capaces, para que puedas crear con agentes potentes sin preocuparte por el coste o la disponibilidad.",
@@ -266,8 +266,6 @@ export const dict = {
   "go.cta.promo": "$5 el primer mes",
   "go.pricing.body":
     "Úsalo con cualquier agente. $5 el primer mes, luego 10 $/mes. Recarga crédito si es necesario. Cancela en cualquier momento.",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6: límite de uso triplicado hasta el 27 de abril",
   "go.graph.free": "Gratis",
   "go.graph.freePill": "Big Pickle y modelos gratuitos",
   "go.graph.go": "Go",
@@ -306,7 +304,7 @@ export const dict = {
   "go.problem.item2": "Límites generosos y acceso fiable",
   "go.problem.item3": "Creado para tantos programadores como sea posible",
   "go.problem.item4":
-    "Incluye GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro y DeepSeek V4 Flash",
+    "Incluye GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro y DeepSeek V4 Flash",
   "go.how.title": "Cómo funciona Go",
   "go.how.body": "Go comienza en $5 el primer mes, luego 10 $/mes. Puedes usarlo con OpenCode o cualquier agente.",
   "go.how.step1.title": "Crear una cuenta",
@@ -331,7 +329,7 @@ export const dict = {
   "go.faq.a2": "Go incluye los modelos que se indican abajo, con límites generosos y acceso confiable.",
   "go.faq.q3": "¿Es Go lo mismo que Zen?",
   "go.faq.a3":
-    "No. Zen es pago por uso, mientras que Go comienza en $5 el primer mes, luego 10 $/mes, con límites generosos y acceso fiable a los modelos de código abierto GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro y DeepSeek V4 Flash.",
+    "No. Zen es pago por uso, mientras que Go comienza en $5 el primer mes, luego 10 $/mes, con límites generosos y acceso fiable a los modelos de código abierto GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro y DeepSeek V4 Flash.",
   "go.faq.q4": "¿Cuánto cuesta Go?",
   "go.faq.a4.p1.beforePricing": "Go cuesta",
   "go.faq.a4.p1.pricingLink": "$5 el primer mes",
@@ -355,7 +353,7 @@ export const dict = {
 
   "go.faq.q9": "¿Cuál es la diferencia entre los modelos gratuitos y Go?",
   "go.faq.a9":
-    "Los modelos gratuitos incluyen Big Pickle más modelos promocionales disponibles en el momento, con una cuota de 200 solicitudes/día. Go incluye GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro y DeepSeek V4 Flash con cuotas de solicitud más altas aplicadas a través de ventanas móviles (5 horas, semanal y mensual), aproximadamente equivalente a 12 $ por 5 horas, 30 $ por semana y 60 $ por mes (los recuentos reales de solicitudes varían según el modelo y el uso).",
+    "Los modelos gratuitos incluyen Big Pickle más modelos promocionales disponibles en el momento, con una cuota de 200 solicitudes/día. Go incluye GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro y DeepSeek V4 Flash con cuotas de solicitud más altas aplicadas a través de ventanas móviles (5 horas, semanal y mensual), aproximadamente equivalente a 12 $ por 5 horas, 30 $ por semana y 60 $ por mes (los recuentos reales de solicitudes varían según el modelo y el uso).",
 
   "zen.api.error.rateLimitExceeded": "Límite de tasa excedido. Por favor, inténtalo de nuevo más tarde.",
   "zen.api.error.modelNotSupported": "Modelo {{model}} no soportado",
diff --git a/packages/console/app/src/i18n/fr.ts b/packages/console/app/src/i18n/fr.ts
index 04e6e3bc62..390025d275 100644
--- a/packages/console/app/src/i18n/fr.ts
+++ b/packages/console/app/src/i18n/fr.ts
@@ -255,7 +255,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | Modèles de code à faible coût pour tous",
   "go.meta.description":
-    "Go commence à $5 pour le premier mois, puis 10 $/mois, avec des limites de requêtes généreuses sur 5 heures pour GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro et DeepSeek V4 Flash.",
+    "Go commence à $5 pour le premier mois, puis 10 $/mois, avec des limites de requêtes généreuses sur 5 heures pour GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro et DeepSeek V4 Flash.",
   "go.hero.title": "Modèles de code à faible coût pour tous",
   "go.hero.body":
     "Go apporte le codage agentique aux programmeurs du monde entier. Offrant des limites généreuses et un accès fiable aux modèles open source les plus capables, pour que vous puissiez construire avec des agents puissants sans vous soucier du coût ou de la disponibilité.",
@@ -267,8 +267,6 @@ export const dict = {
   "go.cta.promo": "$5 le premier mois",
   "go.pricing.body":
     "Utilisez-le avec n'importe quel agent. $5 le premier mois, puis 10 $/mois. Rechargez du crédit si nécessaire. Annulez à tout moment.",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6 : limites d’utilisation triplées jusqu’au 27 avril",
   "go.graph.free": "Gratuit",
   "go.graph.freePill": "Big Pickle et modèles gratuits",
   "go.graph.go": "Go",
@@ -306,7 +304,7 @@ export const dict = {
   "go.problem.item2": "Limites généreuses et accès fiable",
   "go.problem.item3": "Conçu pour autant de programmeurs que possible",
   "go.problem.item4":
-    "Inclut GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro et DeepSeek V4 Flash",
+    "Inclut GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro et DeepSeek V4 Flash",
   "go.how.title": "Comment fonctionne Go",
   "go.how.body":
     "Go commence à $5 pour le premier mois, puis 10 $/mois. Vous pouvez l'utiliser avec OpenCode ou n'importe quel agent.",
@@ -332,7 +330,7 @@ export const dict = {
   "go.faq.a2": "Go inclut les modèles ci-dessous, avec des limites généreuses et un accès fiable.",
   "go.faq.q3": "Est-ce que Go est la même chose que Zen ?",
   "go.faq.a3":
-    "Non. Zen est un paiement à l'utilisation, tandis que Go commence à $5 pour le premier mois, puis 10 $/mois, avec des limites généreuses et un accès fiable aux modèles open source GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro et DeepSeek V4 Flash.",
+    "Non. Zen est un paiement à l'utilisation, tandis que Go commence à $5 pour le premier mois, puis 10 $/mois, avec des limites généreuses et un accès fiable aux modèles open source GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro et DeepSeek V4 Flash.",
   "go.faq.q4": "Combien coûte Go ?",
   "go.faq.a4.p1.beforePricing": "Go coûte",
   "go.faq.a4.p1.pricingLink": "$5 le premier mois",
@@ -355,7 +353,7 @@ export const dict = {
     "Oui, vous pouvez utiliser Go avec n'importe quel agent. Suivez les instructions de configuration dans votre agent de code préféré.",
   "go.faq.q9": "Quelle est la différence entre les modèles gratuits et Go ?",
   "go.faq.a9":
-    "Les modèles gratuits incluent Big Pickle ainsi que des modèles promotionnels disponibles à ce moment-là, avec un quota de 200 requêtes/jour. Go inclut GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro et DeepSeek V4 Flash avec des quotas de requêtes plus élevés appliqués sur des fenêtres glissantes (5 heures, hebdomadaire et mensuelle), à peu près équivalent à 12 $ par 5 heures, 30 $ par semaine et 60 $ par mois (le nombre réel de requêtes varie selon le modèle et l'utilisation).",
+    "Les modèles gratuits incluent Big Pickle ainsi que des modèles promotionnels disponibles à ce moment-là, avec un quota de 200 requêtes/jour. Go inclut GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro et DeepSeek V4 Flash avec des quotas de requêtes plus élevés appliqués sur des fenêtres glissantes (5 heures, hebdomadaire et mensuelle), à peu près équivalent à 12 $ par 5 heures, 30 $ par semaine et 60 $ par mois (le nombre réel de requêtes varie selon le modèle et l'utilisation).",
 
   "zen.api.error.rateLimitExceeded": "Limite de débit dépassée. Veuillez réessayer plus tard.",
   "zen.api.error.modelNotSupported": "Modèle {{model}} non pris en charge",
diff --git a/packages/console/app/src/i18n/it.ts b/packages/console/app/src/i18n/it.ts
index 13f33bfc39..3737186996 100644
--- a/packages/console/app/src/i18n/it.ts
+++ b/packages/console/app/src/i18n/it.ts
@@ -251,7 +251,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | Modelli di coding a basso costo per tutti",
   "go.meta.description":
-    "Go inizia a $5 per il primo mese, poi $10/mese, con generosi limiti di richiesta di 5 ore per GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash.",
+    "Go inizia a $5 per il primo mese, poi $10/mese, con generosi limiti di richiesta di 5 ore per GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash.",
   "go.hero.title": "Modelli di coding a basso costo per tutti",
   "go.hero.body":
     "Go porta il coding agentico ai programmatori di tutto il mondo. Offrendo limiti generosi e un accesso affidabile ai modelli open source più capaci, in modo da poter costruire con agenti potenti senza preoccuparsi dei costi o della disponibilità.",
@@ -263,8 +263,6 @@ export const dict = {
   "go.cta.promo": "$5 il primo mese",
   "go.pricing.body":
     "Usalo con qualsiasi agente. $5 il primo mese, poi $10/mese. Ricarica il credito se necessario. Annulla in qualsiasi momento.",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6: limite d'uso triplicato fino al 27 aprile",
   "go.graph.free": "Gratis",
   "go.graph.freePill": "Big Pickle e modelli gratuiti",
   "go.graph.go": "Go",
@@ -302,7 +300,7 @@ export const dict = {
   "go.problem.item2": "Limiti generosi e accesso affidabile",
   "go.problem.item3": "Costruito per il maggior numero possibile di programmatori",
   "go.problem.item4":
-    "Include GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash",
+    "Include GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash",
   "go.how.title": "Come funziona Go",
   "go.how.body": "Go inizia a $5 per il primo mese, poi $10/mese. Puoi usarlo con OpenCode o qualsiasi agente.",
   "go.how.step1.title": "Crea un account",
@@ -327,7 +325,7 @@ export const dict = {
   "go.faq.a2": "Go include i modelli elencati di seguito, con limiti generosi e accesso affidabile.",
   "go.faq.q3": "Go è lo stesso di Zen?",
   "go.faq.a3":
-    "No. Zen è a consumo, mentre Go inizia a $5 per il primo mese, poi $10/mese, con limiti generosi e accesso affidabile ai modelli open source GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash.",
+    "No. Zen è a consumo, mentre Go inizia a $5 per il primo mese, poi $10/mese, con limiti generosi e accesso affidabile ai modelli open source GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash.",
   "go.faq.q4": "Quanto costa Go?",
   "go.faq.a4.p1.beforePricing": "Go costa",
   "go.faq.a4.p1.pricingLink": "$5 il primo mese",
@@ -351,7 +349,7 @@ export const dict = {
 
   "go.faq.q9": "Qual è la differenza tra i modelli gratuiti e Go?",
   "go.faq.a9":
-    "I modelli gratuiti includono Big Pickle più modelli promozionali disponibili al momento, con una quota di 200 richieste/giorno. Go include GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash con quote di richiesta più elevate applicate su finestre mobili (5 ore, settimanale e mensile), approssimativamente equivalenti a $12 ogni 5 ore, $30 a settimana e $60 al mese (il conteggio effettivo delle richieste varia in base al modello e all'utilizzo).",
+    "I modelli gratuiti includono Big Pickle più modelli promozionali disponibili al momento, con una quota di 200 richieste/giorno. Go include GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro e DeepSeek V4 Flash con quote di richiesta più elevate applicate su finestre mobili (5 ore, settimanale e mensile), approssimativamente equivalenti a $12 ogni 5 ore, $30 a settimana e $60 al mese (il conteggio effettivo delle richieste varia in base al modello e all'utilizzo).",
 
   "zen.api.error.rateLimitExceeded": "Limite di richieste superato. Riprova più tardi.",
   "zen.api.error.modelNotSupported": "Modello {{model}} non supportato",
diff --git a/packages/console/app/src/i18n/ja.ts b/packages/console/app/src/i18n/ja.ts
index 845faebf61..66f3c4a89d 100644
--- a/packages/console/app/src/i18n/ja.ts
+++ b/packages/console/app/src/i18n/ja.ts
@@ -250,7 +250,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | すべての人のための低価格なコーディングモデル",
   "go.meta.description":
-    "Goは最初の月$5、その後$10/月で、GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2-Pro、MiMo-V2-Omni、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro、DeepSeek V4 Flashに対して5時間のゆとりあるリクエスト上限があります。",
+    "Goは最初の月$5、その後$10/月で、GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro、DeepSeek V4 Flashに対して5時間のゆとりあるリクエスト上限があります。",
   "go.hero.title": "すべての人のための低価格なコーディングモデル",
   "go.hero.body":
     "Goは、世界中のプログラマーにエージェント型コーディングをもたらします。最も高性能なオープンソースモデルへの十分な制限と安定したアクセスを提供し、コストや可用性を気にすることなく強力なエージェントで構築できます。",
@@ -262,8 +262,6 @@ export const dict = {
   "go.cta.promo": "初月 $5",
   "go.pricing.body":
     "どのエージェントでも使えます。最初の月$5、その後$10/月。必要に応じてクレジットを追加。いつでもキャンセルできます。",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6、4月27日まで利用上限が3倍に",
   "go.graph.free": "無料",
   "go.graph.freePill": "Big Pickleと無料モデル",
   "go.graph.go": "Go",
@@ -302,7 +300,7 @@ export const dict = {
   "go.problem.item2": "十分な制限と安定したアクセス",
   "go.problem.item3": "できるだけ多くのプログラマーのために構築",
   "go.problem.item4":
-    "GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2-Pro、MiMo-V2-Omni、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro、DeepSeek V4 Flashを含む",
+    "GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro、DeepSeek V4 Flashを含む",
   "go.how.title": "Goの仕組み",
   "go.how.body": "Goは最初の月$5、その後$10/月で始まります。OpenCodeまたは任意のエージェントで使えます。",
   "go.how.step1.title": "アカウントを作成",
@@ -327,7 +325,7 @@ export const dict = {
   "go.faq.a2": "Go には、十分な利用上限と安定したアクセスを備えた、以下のモデルが含まれます。",
   "go.faq.q3": "GoはZenと同じですか？",
   "go.faq.a3":
-    "いいえ。Zenは従量課金制ですが、Goは最初の月$5、その後$10/月で始まり、GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2-Pro、MiMo-V2-Omni、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro、DeepSeek V4 Flashのオープンソースモデルに対して、ゆとりある上限と信頼できるアクセスを提供します。",
+    "いいえ。Zenは従量課金制ですが、Goは最初の月$5、その後$10/月で始まり、GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro、DeepSeek V4 Flashのオープンソースモデルに対して、ゆとりある上限と信頼できるアクセスを提供します。",
   "go.faq.q4": "Goの料金は？",
   "go.faq.a4.p1.beforePricing": "Goは",
   "go.faq.a4.p1.pricingLink": "最初の月$5",
@@ -351,7 +349,7 @@ export const dict = {
 
   "go.faq.q9": "無料モデルとGoの違いは何ですか？",
   "go.faq.a9":
-    "無料モデルにはBig Pickleと、その時点で利用可能なプロモーションモデルが含まれ、1日200リクエストの制限があります。GoにはGLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2-Pro、MiMo-V2-Omni、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro、DeepSeek V4 Flashが含まれ、ローリングウィンドウ（5時間、週間、月間）全体でより高いリクエスト制限が適用されます。これは概算で5時間あたり$12、週間$30、月間$60相当です（実際のリクエスト数はモデルと使用状況により異なります）。",
+    "無料モデルにはBig Pickleと、その時点で利用可能なプロモーションモデルが含まれ、1日200リクエストの制限があります。GoにはGLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro、DeepSeek V4 Flashが含まれ、ローリングウィンドウ（5時間、週間、月間）全体でより高いリクエスト制限が適用されます。これは概算で5時間あたり$12、週間$30、月間$60相当です（実際のリクエスト数はモデルと使用状況により異なります）。",
 
   "zen.api.error.rateLimitExceeded": "レート制限を超えました。後でもう一度お試しください。",
   "zen.api.error.modelNotSupported": "モデル {{model}} はサポートされていません",
diff --git a/packages/console/app/src/i18n/ko.ts b/packages/console/app/src/i18n/ko.ts
index 7efe563a07..04482d35f6 100644
--- a/packages/console/app/src/i18n/ko.ts
+++ b/packages/console/app/src/i18n/ko.ts
@@ -247,7 +247,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | 모두를 위한 저비용 코딩 모델",
   "go.meta.description":
-    "Go는 첫 달 $5, 이후 $10/월로 시작하며, GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, DeepSeek V4 Flash에 대해 넉넉한 5시간 요청 한도를 제공합니다.",
+    "Go는 첫 달 $5, 이후 $10/월로 시작하며, GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, DeepSeek V4 Flash에 대해 넉넉한 5시간 요청 한도를 제공합니다.",
   "go.hero.title": "모두를 위한 저비용 코딩 모델",
   "go.hero.body":
     "Go는 전 세계 프로그래머들에게 에이전트 코딩을 제공합니다. 가장 유능한 오픈 소스 모델에 대한 넉넉한 한도와 안정적인 액세스를 제공하므로, 비용이나 가용성 걱정 없이 강력한 에이전트로 빌드할 수 있습니다.",
@@ -259,8 +259,6 @@ export const dict = {
   "go.cta.promo": "첫 달 $5",
   "go.pricing.body":
     "어떤 에이전트와도 사용할 수 있습니다. 첫 달 $5, 이후 $10/월. 필요하면 크레딧을 충전하세요. 언제든지 취소할 수 있습니다.",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6, 4월 27일까지 사용 한도 3배 확대",
   "go.graph.free": "무료",
   "go.graph.freePill": "Big Pickle 및 무료 모델",
   "go.graph.go": "Go",
@@ -299,7 +297,7 @@ export const dict = {
   "go.problem.item2": "넉넉한 한도와 안정적인 액세스",
   "go.problem.item3": "가능한 한 많은 프로그래머를 위해 제작됨",
   "go.problem.item4":
-    "GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, DeepSeek V4 Flash 포함",
+    "GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, DeepSeek V4 Flash 포함",
   "go.how.title": "Go 작동 방식",
   "go.how.body": "Go는 첫 달 $5, 이후 $10/월로 시작합니다. OpenCode 또는 어떤 에이전트와도 함께 사용할 수 있습니다.",
   "go.how.step1.title": "계정 생성",
@@ -323,7 +321,7 @@ export const dict = {
   "go.faq.a2": "Go에는 넉넉한 한도와 안정적인 액세스를 제공하는 아래 모델이 포함됩니다.",
   "go.faq.q3": "Go는 Zen과 같은가요?",
   "go.faq.a3":
-    "아니요. Zen은 종량제인 반면, Go는 첫 달 $5, 이후 $10/월로 시작하며, GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, DeepSeek V4 Flash 오픈 소스 모델에 대한 넉넉한 한도와 안정적인 액세스를 제공합니다.",
+    "아니요. Zen은 종량제인 반면, Go는 첫 달 $5, 이후 $10/월로 시작하며, GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, DeepSeek V4 Flash 오픈 소스 모델에 대한 넉넉한 한도와 안정적인 액세스를 제공합니다.",
   "go.faq.q4": "Go 비용은 얼마인가요?",
   "go.faq.a4.p1.beforePricing": "Go 비용은",
   "go.faq.a4.p1.pricingLink": "첫 달 $5",
@@ -346,7 +344,7 @@ export const dict = {
 
   "go.faq.q9": "무료 모델과 Go의 차이점은 무엇인가요?",
   "go.faq.a9":
-    "무료 모델에는 Big Pickle과 당시 사용 가능한 프로모션 모델이 포함되며, 하루 200회 요청 할당량이 적용됩니다. Go는 GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, DeepSeek V4 Flash를 포함하며, 롤링 윈도우(5시간, 주간, 월간)에 걸쳐 더 높은 요청 할당량을 적용합니다. 이는 대략 5시간당 $12, 주당 $30, 월 $60에 해당합니다(실제 요청 수는 모델 및 사용량에 따라 다름).",
+    "무료 모델에는 Big Pickle과 당시 사용 가능한 프로모션 모델이 포함되며, 하루 200회 요청 할당량이 적용됩니다. Go는 GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro, DeepSeek V4 Flash를 포함하며, 롤링 윈도우(5시간, 주간, 월간)에 걸쳐 더 높은 요청 할당량을 적용합니다. 이는 대략 5시간당 $12, 주당 $30, 월 $60에 해당합니다(실제 요청 수는 모델 및 사용량에 따라 다름).",
 
   "zen.api.error.rateLimitExceeded": "속도 제한을 초과했습니다. 나중에 다시 시도해 주세요.",
   "zen.api.error.modelNotSupported": "{{model}} 모델은 지원되지 않습니다",
diff --git a/packages/console/app/src/i18n/no.ts b/packages/console/app/src/i18n/no.ts
index 8948e158b0..31200d3edd 100644
--- a/packages/console/app/src/i18n/no.ts
+++ b/packages/console/app/src/i18n/no.ts
@@ -251,7 +251,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | Rimelige kodemodeller for alle",
   "go.meta.description":
-    "Go starter på $5 for den første måneden, deretter $10/måned, med sjenerøse 5-timers forespørselsgrenser for GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash.",
+    "Go starter på $5 for den første måneden, deretter $10/måned, med sjenerøse 5-timers forespørselsgrenser for GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash.",
   "go.hero.title": "Rimelige kodemodeller for alle",
   "go.hero.body":
     "Go bringer agent-koding til programmerere over hele verden. Med rause grenser og pålitelig tilgang til de mest kapable åpen kildekode-modellene, kan du bygge med kraftige agenter uten å bekymre deg for kostnader eller tilgjengelighet.",
@@ -263,8 +263,6 @@ export const dict = {
   "go.cta.promo": "$5 første måned",
   "go.pricing.body":
     "Bruk med hvilken som helst agent. $5 første måned, deretter $10/måned. Fyll på kreditt ved behov. Avslutt når som helst.",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6: bruksgrensen er tredoblet til 27. april",
   "go.graph.free": "Gratis",
   "go.graph.freePill": "Big Pickle og gratis modeller",
   "go.graph.go": "Go",
@@ -302,7 +300,7 @@ export const dict = {
   "go.problem.item2": "Rause grenser og pålitelig tilgang",
   "go.problem.item3": "Bygget for så mange programmerere som mulig",
   "go.problem.item4":
-    "Inkluderer GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash",
+    "Inkluderer GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash",
   "go.how.title": "Hvordan Go fungerer",
   "go.how.body":
     "Go starter på $5 for den første måneden, deretter $10/måned. Du kan bruke det med OpenCode eller hvilken som helst agent.",
@@ -328,7 +326,7 @@ export const dict = {
   "go.faq.a2": "Go inkluderer modellene nedenfor, med høye grenser og pålitelig tilgang.",
   "go.faq.q3": "Er Go det samme som Zen?",
   "go.faq.a3":
-    "Nei. Zen er betaling etter bruk, mens Go starter på $5 for den første måneden, deretter $10/måned, med sjenerøse grenser og pålitelig tilgang til åpen kildekode-modellene GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash.",
+    "Nei. Zen er betaling etter bruk, mens Go starter på $5 for den første måneden, deretter $10/måned, med sjenerøse grenser og pålitelig tilgang til åpen kildekode-modellene GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash.",
   "go.faq.q4": "Hva koster Go?",
   "go.faq.a4.p1.beforePricing": "Go koster",
   "go.faq.a4.p1.pricingLink": "$5 første måned",
@@ -352,7 +350,7 @@ export const dict = {
 
   "go.faq.q9": "Hva er forskjellen mellom gratis modeller og Go?",
   "go.faq.a9":
-    "Gratis modeller inkluderer Big Pickle pluss kampanjemodeller tilgjengelig på det tidspunktet, med en kvote på 200 forespørsler/dag. Go inkluderer GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash med høyere kvoter håndhevet over rullerende vinduer (5 timer, ukentlig og månedlig), omtrent tilsvarende $12 per 5 timer, $30 per uke og $60 per måned (faktiske forespørselsantall varierer etter modell og bruk).",
+    "Gratis modeller inkluderer Big Pickle pluss kampanjemodeller tilgjengelig på det tidspunktet, med en kvote på 200 forespørsler/dag. Go inkluderer GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro og DeepSeek V4 Flash med høyere kvoter håndhevet over rullerende vinduer (5 timer, ukentlig og månedlig), omtrent tilsvarende $12 per 5 timer, $30 per uke og $60 per måned (faktiske forespørselsantall varierer etter modell og bruk).",
 
   "zen.api.error.rateLimitExceeded": "Rate limit overskredet. Vennligst prøv igjen senere.",
   "zen.api.error.modelNotSupported": "Modell {{model}} støttes ikke",
diff --git a/packages/console/app/src/i18n/pl.ts b/packages/console/app/src/i18n/pl.ts
index f879ed7057..50d904bc56 100644
--- a/packages/console/app/src/i18n/pl.ts
+++ b/packages/console/app/src/i18n/pl.ts
@@ -252,7 +252,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | Niskokosztowe modele do kodowania dla każdego",
   "go.meta.description":
-    "Go zaczyna się od $5 za pierwszy miesiąc, potem $10/miesiąc, z hojnymi 5-godzinnymi limitami zapytań dla GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro i DeepSeek V4 Flash.",
+    "Go zaczyna się od $5 za pierwszy miesiąc, potem $10/miesiąc, z hojnymi 5-godzinnymi limitami zapytań dla GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro i DeepSeek V4 Flash.",
   "go.hero.title": "Niskokosztowe modele do kodowania dla każdego",
   "go.hero.body":
     "Go udostępnia programowanie z agentami programistom na całym świecie. Oferuje hojne limity i niezawodny dostęp do najzdolniejszych modeli open source, dzięki czemu możesz budować za pomocą potężnych agentów, nie martwiąc się o koszty czy dostępność.",
@@ -264,8 +264,6 @@ export const dict = {
   "go.cta.promo": "$5 pierwszy miesiąc",
   "go.pricing.body":
     "Używaj z dowolnym agentem. $5 za pierwszy miesiąc, potem $10/miesiąc. Doładuj konto w razie potrzeby. Anuluj w dowolnym momencie.",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6: limit użycia zwiększony 3× do 27 kwietnia",
   "go.graph.free": "Darmowe",
   "go.graph.freePill": "Big Pickle i darmowe modele",
   "go.graph.go": "Go",
@@ -303,7 +301,7 @@ export const dict = {
   "go.problem.item2": "Hojne limity i niezawodny dostęp",
   "go.problem.item3": "Stworzony dla jak największej liczby programistów",
   "go.problem.item4":
-    "Zawiera GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro i DeepSeek V4 Flash",
+    "Zawiera GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro i DeepSeek V4 Flash",
   "go.how.title": "Jak działa Go",
   "go.how.body":
     "Go zaczyna się od $5 za pierwszy miesiąc, potem $10/miesiąc. Możesz go używać z OpenCode lub dowolnym agentem.",
@@ -329,7 +327,7 @@ export const dict = {
   "go.faq.a2": "Go obejmuje poniższe modele z wysokimi limitami i niezawodnym dostępem.",
   "go.faq.q3": "Czy Go to to samo co Zen?",
   "go.faq.a3":
-    "Nie. Zen to model płatności za użycie, podczas gdy Go zaczyna się od $5 za pierwszy miesiąc, potem $10/miesiąc, z hojnymi limitami i niezawodnym dostępem do modeli open source GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro i DeepSeek V4 Flash.",
+    "Nie. Zen to model płatności za użycie, podczas gdy Go zaczyna się od $5 za pierwszy miesiąc, potem $10/miesiąc, z hojnymi limitami i niezawodnym dostępem do modeli open source GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro i DeepSeek V4 Flash.",
   "go.faq.q4": "Ile kosztuje Go?",
   "go.faq.a4.p1.beforePricing": "Go kosztuje",
   "go.faq.a4.p1.pricingLink": "$5 za pierwszy miesiąc",
@@ -353,7 +351,7 @@ export const dict = {
 
   "go.faq.q9": "Jaka jest różnica między darmowymi modelami a Go?",
   "go.faq.a9":
-    "Darmowe modele obejmują Big Pickle oraz modele promocyjne dostępne w danym momencie, z limitem 200 zapytań/dzień. Go zawiera GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro i DeepSeek V4 Flash z wyższymi limitami zapytań egzekwowanymi w oknach kroczących (5-godzinnych, tygodniowych i miesięcznych), w przybliżeniu równoważnymi $12 na 5 godzin, $30 tygodniowo i $60 miesięcznie (rzeczywista liczba zapytań zależy od modelu i użycia).",
+    "Darmowe modele obejmują Big Pickle oraz modele promocyjne dostępne w danym momencie, z limitem 200 zapytań/dzień. Go zawiera GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro i DeepSeek V4 Flash z wyższymi limitami zapytań egzekwowanymi w oknach kroczących (5-godzinnych, tygodniowych i miesięcznych), w przybliżeniu równoważnymi $12 na 5 godzin, $30 tygodniowo i $60 miesięcznie (rzeczywista liczba zapytań zależy od modelu i użycia).",
 
   "zen.api.error.rateLimitExceeded": "Przekroczono limit zapytań. Spróbuj ponownie później.",
   "zen.api.error.modelNotSupported": "Model {{model}} nie jest obsługiwany",
diff --git a/packages/console/app/src/i18n/ru.ts b/packages/console/app/src/i18n/ru.ts
index 9ba36d2208..651309fc95 100644
--- a/packages/console/app/src/i18n/ru.ts
+++ b/packages/console/app/src/i18n/ru.ts
@@ -255,7 +255,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | Недорогие модели для кодинга для всех",
   "go.meta.description":
-    "Go начинается с $5 за первый месяц, затем $10/месяц, с щедрыми лимитами запросов за 5 часов для GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro и DeepSeek V4 Flash.",
+    "Go начинается с $5 за первый месяц, затем $10/месяц, с щедрыми лимитами запросов за 5 часов для GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro и DeepSeek V4 Flash.",
   "go.hero.title": "Недорогие модели для кодинга для всех",
   "go.hero.body":
     "Go открывает доступ к агентам-программистам разработчикам по всему миру. Предлагая щедрые лимиты и надежный доступ к наиболее способным моделям с открытым исходным кодом, вы можете создавать проекты с мощными агентами, не беспокоясь о затратах или доступности.",
@@ -267,8 +267,6 @@ export const dict = {
   "go.cta.promo": "$5 первый месяц",
   "go.pricing.body":
     "Используйте с любым агентом. $5 за первый месяц, затем $10/месяц. Пополняйте баланс при необходимости. Отменить можно в любое время.",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6: лимит использования увеличен в 3 раза до 27 апреля",
   "go.graph.free": "Бесплатно",
   "go.graph.freePill": "Big Pickle и бесплатные модели",
   "go.graph.go": "Go",
@@ -307,7 +305,7 @@ export const dict = {
   "go.problem.item2": "Щедрые лимиты и надежный доступ",
   "go.problem.item3": "Создан для максимального числа программистов",
   "go.problem.item4":
-    "Включает GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro и DeepSeek V4 Flash",
+    "Включает GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro и DeepSeek V4 Flash",
   "go.how.title": "Как работает Go",
   "go.how.body":
     "Go начинается с $5 за первый месяц, затем $10/месяц. Вы можете использовать его с OpenCode или любым агентом.",
@@ -333,7 +331,7 @@ export const dict = {
   "go.faq.a2": "Go включает перечисленные ниже модели с щедрыми лимитами и надежным доступом.",
   "go.faq.q3": "Go — это то же самое, что и Zen?",
   "go.faq.a3":
-    "Нет. Zen - это оплата по мере использования, в то время как Go начинается с $5 за первый месяц, затем $10/месяц, с щедрыми лимитами и надежным доступом к моделям с открытым исходным кодом GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro и DeepSeek V4 Flash.",
+    "Нет. Zen - это оплата по мере использования, в то время как Go начинается с $5 за первый месяц, затем $10/месяц, с щедрыми лимитами и надежным доступом к моделям с открытым исходным кодом GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro и DeepSeek V4 Flash.",
   "go.faq.q4": "Сколько стоит Go?",
   "go.faq.a4.p1.beforePricing": "Go стоит",
   "go.faq.a4.p1.pricingLink": "$5 за первый месяц",
@@ -357,7 +355,7 @@ export const dict = {
 
   "go.faq.q9": "В чем разница между бесплатными моделями и Go?",
   "go.faq.a9":
-    "Бесплатные модели включают Big Pickle плюс промо-модели, доступные на данный момент, с квотой 200 запросов/день. Go включает GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro и DeepSeek V4 Flash с более высокими квотами запросов, применяемыми в скользящих окнах (5 часов, неделя и месяц), что примерно эквивалентно $12 за 5 часов, $30 в неделю и $60 в месяц (фактическое количество запросов зависит от модели и использования).",
+    "Бесплатные модели включают Big Pickle плюс промо-модели, доступные на данный момент, с квотой 200 запросов/день. Go включает GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro и DeepSeek V4 Flash с более высокими квотами запросов, применяемыми в скользящих окнах (5 часов, неделя и месяц), что примерно эквивалентно $12 за 5 часов, $30 в неделю и $60 в месяц (фактическое количество запросов зависит от модели и использования).",
 
   "zen.api.error.rateLimitExceeded": "Превышен лимит запросов. Пожалуйста, попробуйте позже.",
   "zen.api.error.modelNotSupported": "Модель {{model}} не поддерживается",
diff --git a/packages/console/app/src/i18n/th.ts b/packages/console/app/src/i18n/th.ts
index 01b2b19c39..42c9e455fd 100644
--- a/packages/console/app/src/i18n/th.ts
+++ b/packages/console/app/src/i18n/th.ts
@@ -250,7 +250,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | โมเดลเขียนโค้ดราคาประหยัดสำหรับทุกคน",
   "go.meta.description":
-    "Go เริ่มต้นที่ $5 สำหรับเดือนแรก จากนั้น $10/เดือน พร้อมขีดจำกัดคำขอ 5 ชั่วโมงที่เอื้อเฟื้อสำหรับ GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro และ DeepSeek V4 Flash",
+    "Go เริ่มต้นที่ $5 สำหรับเดือนแรก จากนั้น $10/เดือน พร้อมขีดจำกัดคำขอ 5 ชั่วโมงที่เอื้อเฟื้อสำหรับ GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro และ DeepSeek V4 Flash",
   "go.hero.title": "โมเดลเขียนโค้ดราคาประหยัดสำหรับทุกคน",
   "go.hero.body":
     "Go นำการเขียนโค้ดแบบเอเจนต์มาสู่นักเขียนโปรแกรมทั่วโลก เสนอขีดจำกัดที่กว้างขวางและการเข้าถึงโมเดลโอเพนซอร์สที่มีความสามารถสูงสุดได้อย่างน่าเชื่อถือ เพื่อให้คุณสามารถสร้างสรรค์ด้วยเอเจนต์ที่ทรงพลังโดยไม่ต้องกังวลเรื่องค่าใช้จ่ายหรือความพร้อมใช้งาน",
@@ -261,8 +261,6 @@ export const dict = {
   "go.cta.price": "$10/เดือน",
   "go.cta.promo": "$5 เดือนแรก",
   "go.pricing.body": "ใช้กับเอเจนต์ใดก็ได้ $5 ในเดือนแรก จากนั้น $10/เดือน เติมเครดิตหากจำเป็น ยกเลิกได้ตลอดเวลา",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6 โควตาการใช้งานเพิ่มเป็น 3 เท่า ถึง 27 เม.ย.",
   "go.graph.free": "ฟรี",
   "go.graph.freePill": "Big Pickle และโมเดลฟรี",
   "go.graph.go": "Go",
@@ -300,7 +298,7 @@ export const dict = {
   "go.problem.item2": "ขีดจำกัดที่กว้างขวางและการเข้าถึงที่เชื่อถือได้",
   "go.problem.item3": "สร้างขึ้นเพื่อโปรแกรมเมอร์จำนวนมากที่สุดเท่าที่จะเป็นไปได้",
   "go.problem.item4":
-    "รวมถึง GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro และ DeepSeek V4 Flash",
+    "รวมถึง GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro และ DeepSeek V4 Flash",
   "go.how.title": "Go ทำงานอย่างไร",
   "go.how.body": "Go เริ่มต้นที่ $5 สำหรับเดือนแรก จากนั้น $10/เดือน คุณสามารถใช้กับ OpenCode หรือเอเจนต์ใดก็ได้",
   "go.how.step1.title": "สร้างบัญชี",
@@ -325,7 +323,7 @@ export const dict = {
   "go.faq.a2": "Go รวมโมเดลด้านล่างนี้ พร้อมขีดจำกัดที่มากและการเข้าถึงที่เชื่อถือได้",
   "go.faq.q3": "Go เหมือนกับ Zen หรือไม่?",
   "go.faq.a3":
-    "ไม่ Zen เป็นแบบจ่ายตามการใช้งาน ในขณะที่ Go เริ่มต้นที่ $5 สำหรับเดือนแรก จากนั้น $10/เดือน พร้อมขีดจำกัดที่เอื้อเฟื้อและการเข้าถึงโมเดลโอเพนซอร์ส GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro และ DeepSeek V4 Flash อย่างเชื่อถือได้",
+    "ไม่ Zen เป็นแบบจ่ายตามการใช้งาน ในขณะที่ Go เริ่มต้นที่ $5 สำหรับเดือนแรก จากนั้น $10/เดือน พร้อมขีดจำกัดที่เอื้อเฟื้อและการเข้าถึงโมเดลโอเพนซอร์ส GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro และ DeepSeek V4 Flash อย่างเชื่อถือได้",
   "go.faq.q4": "Go ราคาเท่าไหร่?",
   "go.faq.a4.p1.beforePricing": "Go ราคา",
   "go.faq.a4.p1.pricingLink": "$5 เดือนแรก",
@@ -348,7 +346,7 @@ export const dict = {
 
   "go.faq.q9": "ความแตกต่างระหว่างโมเดลฟรีและ Go คืออะไร?",
   "go.faq.a9":
-    "โมเดลฟรีรวมถึง Big Pickle บวกกับโมเดลโปรโมชั่นที่มีให้ในขณะนั้น ด้วยโควต้า 200 คำขอ/วัน Go รวมถึง GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro และ DeepSeek V4 Flash ที่มีโควต้าคำขอสูงกว่า ซึ่งบังคับใช้ผ่านช่วงเวลาหมุนเวียน (5 ชั่วโมง, รายสัปดาห์ และรายเดือน) เทียบเท่าประมาณ $12 ต่อ 5 ชั่วโมง, $30 ต่อสัปดาห์ และ $60 ต่อเดือน (จำนวนคำขอจริงจะแตกต่างกันไปตามโมเดลและการใช้งาน)",
+    "โมเดลฟรีรวมถึง Big Pickle บวกกับโมเดลโปรโมชั่นที่มีให้ในขณะนั้น ด้วยโควต้า 200 คำขอ/วัน Go รวมถึง GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro และ DeepSeek V4 Flash ที่มีโควต้าคำขอสูงกว่า ซึ่งบังคับใช้ผ่านช่วงเวลาหมุนเวียน (5 ชั่วโมง, รายสัปดาห์ และรายเดือน) เทียบเท่าประมาณ $12 ต่อ 5 ชั่วโมง, $30 ต่อสัปดาห์ และ $60 ต่อเดือน (จำนวนคำขอจริงจะแตกต่างกันไปตามโมเดลและการใช้งาน)",
 
   "zen.api.error.rateLimitExceeded": "เกินขีดจำกัดอัตราการใช้งาน กรุณาลองใหม่ในภายหลัง",
   "zen.api.error.modelNotSupported": "ไม่รองรับโมเดล {{model}}",
diff --git a/packages/console/app/src/i18n/tr.ts b/packages/console/app/src/i18n/tr.ts
index 0345277b87..64380db375 100644
--- a/packages/console/app/src/i18n/tr.ts
+++ b/packages/console/app/src/i18n/tr.ts
@@ -253,7 +253,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | Herkes için düşük maliyetli kodlama modelleri",
   "go.meta.description":
-    "Go ilk ay $5, sonrasında ayda 10$ fiyatıyla başlar; GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro ve DeepSeek V4 Flash için cömert 5 saatlik istek limitleri sunar.",
+    "Go ilk ay $5, sonrasında ayda 10$ fiyatıyla başlar; GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro ve DeepSeek V4 Flash için cömert 5 saatlik istek limitleri sunar.",
   "go.hero.title": "Herkes için düşük maliyetli kodlama modelleri",
   "go.hero.body":
     "Go, dünya çapındaki programcılara ajan tabanlı kodlama getiriyor. En yetenekli açık kaynaklı modellere cömert limitler ve güvenilir erişim sunarak, maliyet veya erişilebilirlik konusunda endişelenmeden güçlü ajanlarla geliştirme yapmanızı sağlar.",
@@ -265,8 +265,6 @@ export const dict = {
   "go.cta.promo": "İlk ay $5",
   "go.pricing.body":
     "Herhangi bir ajanla kullanın. İlk ay $5, sonrasında ayda 10$. Gerekirse kredi yükleyin. İstediğiniz zaman iptal edin.",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6: kullanım limiti 27 Nisan'a kadar 3 katına çıktı",
   "go.graph.free": "Ücretsiz",
   "go.graph.freePill": "Big Pickle ve ücretsiz modeller",
   "go.graph.go": "Go",
@@ -305,7 +303,7 @@ export const dict = {
   "go.problem.item2": "Cömert limitler ve güvenilir erişim",
   "go.problem.item3": "Mümkün olduğunca çok programcı için geliştirildi",
   "go.problem.item4":
-    "GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro ve DeepSeek V4 Flash içerir",
+    "GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro ve DeepSeek V4 Flash içerir",
   "go.how.title": "Go nasıl çalışır?",
   "go.how.body":
     "Go ilk ay $5, sonrasında ayda 10$ fiyatıyla başlar. OpenCode veya herhangi bir ajanla kullanabilirsiniz.",
@@ -331,7 +329,7 @@ export const dict = {
   "go.faq.a2": "Go, aşağıda listelenen modelleri cömert limitler ve güvenilir erişimle sunar.",
   "go.faq.q3": "Go, Zen ile aynı mı?",
   "go.faq.a3":
-    "Hayır. Zen kullandıkça öde modelidir, Go ise ilk ay $5, sonrasında ayda 10$ fiyatıyla başlar; GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro ve DeepSeek V4 Flash açık kaynak modellerine cömert limitler ve güvenilir erişim sunar.",
+    "Hayır. Zen kullandıkça öde modelidir, Go ise ilk ay $5, sonrasında ayda 10$ fiyatıyla başlar; GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro ve DeepSeek V4 Flash açık kaynak modellerine cömert limitler ve güvenilir erişim sunar.",
   "go.faq.q4": "Go ne kadar?",
   "go.faq.a4.p1.beforePricing": "Go'nun maliyeti",
   "go.faq.a4.p1.pricingLink": "İlk ay $5",
@@ -355,7 +353,7 @@ export const dict = {
 
   "go.faq.q9": "Ücretsiz modeller ve Go arasındaki fark nedir?",
   "go.faq.a9":
-    "Ücretsiz modeller, günlük 200 istek kotası ile Big Pickle ve o sırada mevcut olan promosyonel modelleri içerir. Go ise GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro ve DeepSeek V4 Flash modellerini; yuvarlanan pencereler (5 saatlik, haftalık ve aylık) üzerinden uygulanan daha yüksek istek kotalarıyla içerir. Bu kotalar kabaca her 5 saatte 12$, haftada 30$ ve ayda 60$ değerine eşdeğerdir (gerçek istek sayıları modele ve kullanıma göre değişir).",
+    "Ücretsiz modeller, günlük 200 istek kotası ile Big Pickle ve o sırada mevcut olan promosyonel modelleri içerir. Go ise GLM-5.1, GLM-5, Kimi K2.5, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5, MiniMax M2.7, DeepSeek V4 Pro ve DeepSeek V4 Flash modellerini; yuvarlanan pencereler (5 saatlik, haftalık ve aylık) üzerinden uygulanan daha yüksek istek kotalarıyla içerir. Bu kotalar kabaca her 5 saatte 12$, haftada 30$ ve ayda 60$ değerine eşdeğerdir (gerçek istek sayıları modele ve kullanıma göre değişir).",
 
   "zen.api.error.rateLimitExceeded": "İstek limiti aşıldı. Lütfen daha sonra tekrar deneyin.",
   "zen.api.error.modelNotSupported": "{{model}} modeli desteklenmiyor",
diff --git a/packages/console/app/src/i18n/zh.ts b/packages/console/app/src/i18n/zh.ts
index b9300cc87e..3b104cca6d 100644
--- a/packages/console/app/src/i18n/zh.ts
+++ b/packages/console/app/src/i18n/zh.ts
@@ -241,7 +241,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | 人人可用的低成本编程模型",
   "go.meta.description":
-    "Go 首月 $5，之后 $10/月，提供对 GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2-Pro、MiMo-V2-Omni、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 和 DeepSeek V4 Flash 的 5 小时充裕请求额度。",
+    "Go 首月 $5，之后 $10/月，提供对 GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 和 DeepSeek V4 Flash 的 5 小时充裕请求额度。",
   "go.hero.title": "人人可用的低成本编程模型",
   "go.hero.body":
     "Go 将代理编程带给全世界的程序员。提供充裕的限额和对最强大的开源模型的可靠访问，让您可以利用强大的代理进行构建，而无需担心成本或可用性。",
@@ -252,8 +252,6 @@ export const dict = {
   "go.cta.price": "$10/月",
   "go.cta.promo": "首月 $5",
   "go.pricing.body": "可配合任何代理使用。首月 $5，之后 $10/月。如有需要可充值。随时取消。",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6 使用额度提升至 3 倍，限时至 4 月 27 日",
   "go.graph.free": "免费",
   "go.graph.freePill": "Big Pickle 和免费模型",
   "go.graph.go": "Go",
@@ -291,7 +289,7 @@ export const dict = {
   "go.problem.item2": "充裕的限额和可靠的访问",
   "go.problem.item3": "为尽可能多的程序员打造",
   "go.problem.item4":
-    "包含 GLM-5.1, GLM-5, Kimi K2.5、Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 和 DeepSeek V4 Flash",
+    "包含 GLM-5.1, GLM-5, Kimi K2.5、Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 和 DeepSeek V4 Flash",
   "go.how.title": "Go 如何工作",
   "go.how.body": "Go 起价为首月 $5，之后 $10/月。您可以将其与 OpenCode 或任何代理搭配使用。",
   "go.how.step1.title": "创建账户",
@@ -313,7 +311,7 @@ export const dict = {
   "go.faq.a2": "Go 包含下方列出的模型，提供充足的限额和可靠的访问。",
   "go.faq.q3": "Go 和 Zen 一样吗？",
   "go.faq.a3":
-    "不。Zen 是按量付费，而 Go 首月 $5，之后 $10/月，提供充裕的额度，并可可靠地访问 GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2-Pro、MiMo-V2-Omni、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 和 DeepSeek V4 Flash 等开源模型。",
+    "不。Zen 是按量付费，而 Go 首月 $5，之后 $10/月，提供充裕的额度，并可可靠地访问 GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 和 DeepSeek V4 Flash 等开源模型。",
   "go.faq.q4": "Go 多少钱？",
   "go.faq.a4.p1.beforePricing": "Go 费用为",
   "go.faq.a4.p1.pricingLink": "首月 $5",
@@ -335,7 +333,7 @@ export const dict = {
 
   "go.faq.q9": "免费模型和 Go 之间的区别是什么？",
   "go.faq.a9":
-    "免费模型包含 Big Pickle 加上当时可用的促销模型，每天有 200 次请求的配额。Go 包含 GLM-5.1, GLM-5, Kimi K2.5、Kimi K2.6, MiMo-V2-Pro, MiMo-V2-Omni, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 和 DeepSeek V4 Flash，并在滚动窗口（5 小时、每周和每月）内执行更高的请求配额，大致相当于每 5 小时 $12、每周 $30 和每月 $60（实际请求计数因模型和使用情况而异）。",
+    "免费模型包含 Big Pickle 加上当时可用的促销模型，每天有 200 次请求的配额。Go 包含 GLM-5.1, GLM-5, Kimi K2.5、Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.5 Plus, Qwen3.6 Plus, MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 和 DeepSeek V4 Flash，并在滚动窗口（5 小时、每周和每月）内执行更高的请求配额，大致相当于每 5 小时 $12、每周 $30 和每月 $60（实际请求计数因模型和使用情况而异）。",
 
   "zen.api.error.rateLimitExceeded": "超出速率限制。请稍后重试。",
   "zen.api.error.modelNotSupported": "不支持模型 {{model}}",
diff --git a/packages/console/app/src/i18n/zht.ts b/packages/console/app/src/i18n/zht.ts
index f129a99d02..a4d5512da4 100644
--- a/packages/console/app/src/i18n/zht.ts
+++ b/packages/console/app/src/i18n/zht.ts
@@ -241,7 +241,7 @@ export const dict = {
 
   "go.title": "OpenCode Go | 低成本全民編碼模型",
   "go.meta.description":
-    "Go 首月 $5，之後 $10/月，提供對 GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2-Pro、MiMo-V2-Omni、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 和 DeepSeek V4 Flash 的 5 小時充裕請求額度。",
+    "Go 首月 $5，之後 $10/月，提供對 GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 和 DeepSeek V4 Flash 的 5 小時充裕請求額度。",
   "go.hero.title": "低成本全民編碼模型",
   "go.hero.body":
     "Go 將代理編碼帶給全世界的程式設計師。提供寬裕的限額以及對最強大開源模型的穩定存取，讓你可以使用強大的代理進行構建，而無需擔心成本或可用性。",
@@ -252,8 +252,6 @@ export const dict = {
   "go.cta.price": "$10/月",
   "go.cta.promo": "首月 $5",
   "go.pricing.body": "可搭配任何代理使用。首月 $5，之後 $10/月。如有需要可儲值。隨時取消。",
-  "go.banner.badge": "3x",
-  "go.banner.text": "Kimi K2.6 使用額度提升至 3 倍，限時至 4 月 27 日",
   "go.graph.free": "免費",
   "go.graph.freePill": "Big Pickle 與免費模型",
   "go.graph.go": "Go",
@@ -291,7 +289,7 @@ export const dict = {
   "go.problem.item2": "寬裕的限額與穩定存取",
   "go.problem.item3": "專為盡可能多的程式設計師打造",
   "go.problem.item4":
-    "包含 GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2-Pro、MiMo-V2-Omni、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 與 DeepSeek V4 Flash",
+    "包含 GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 與 DeepSeek V4 Flash",
   "go.how.title": "Go 如何運作",
   "go.how.body": "Go 起價為首月 $5，之後 $10/月。您可以將其與 OpenCode 或任何代理搭配使用。",
   "go.how.step1.title": "建立帳號",
@@ -313,7 +311,7 @@ export const dict = {
   "go.faq.a2": "Go 包含下方列出的模型，提供充足的額度與穩定的存取。",
   "go.faq.q3": "Go 與 Zen 一樣嗎？",
   "go.faq.a3":
-    "不。Zen 是按量付費，而 Go 首月 $5，之後 $10/月，提供充裕的額度，並可可靠地存取 GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2-Pro、MiMo-V2-Omni、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 和 DeepSeek V4 Flash 等開源模型。",
+    "不。Zen 是按量付費，而 Go 首月 $5，之後 $10/月，提供充裕的額度，並可可靠地存取 GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 和 DeepSeek V4 Flash 等開源模型。",
   "go.faq.q4": "Go 費用是多少？",
   "go.faq.a4.p1.beforePricing": "Go 費用為",
   "go.faq.a4.p1.pricingLink": "首月 $5",
@@ -335,7 +333,7 @@ export const dict = {
 
   "go.faq.q9": "免費模型與 Go 有什麼區別？",
   "go.faq.a9":
-    "免費模型包括 Big Pickle 以及當時可用的促銷模型，配額為 200 次請求/天。Go 包括 GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2-Pro、MiMo-V2-Omni、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 與 DeepSeek V4 Flash，並在滾動視窗（5 小時、每週和每月）內執行更高的請求配額，大約相當於每 5 小時 $12、每週 $30 和每月 $60（實際請求數因模型和使用情況而異）。",
+    "免費模型包括 Big Pickle 以及當時可用的促銷模型，配額為 200 次請求/天。Go 包括 GLM-5.1、GLM-5、Kimi K2.5、Kimi K2.6、MiMo-V2.5-Pro、MiMo-V2.5、Qwen3.5 Plus、Qwen3.6 Plus、MiniMax M2.5、MiniMax M2.7、DeepSeek V4 Pro 與 DeepSeek V4 Flash，並在滾動視窗（5 小時、每週和每月）內執行更高的請求配額，大約相當於每 5 小時 $12、每週 $30 和每月 $60（實際請求數因模型和使用情況而異）。",
 
   "zen.api.error.rateLimitExceeded": "超出頻率限制。請稍後再試。",
   "zen.api.error.modelNotSupported": "不支援模型 {{model}}",
diff --git a/packages/console/app/src/routes/go/index.css b/packages/console/app/src/routes/go/index.css
index de8dce4724..25ae00e5f8 100644
--- a/packages/console/app/src/routes/go/index.css
+++ b/packages/console/app/src/routes/go/index.css
@@ -326,37 +326,6 @@ body {
     }
   }
 
-  [data-component="desktop-app-banner"] {
-    display: flex;
-    align-items: center;
-    gap: 12px;
-    margin-bottom: 32px;
-
-    [data-slot="badge"] {
-      background: var(--color-background-strong);
-      color: var(--color-text-inverted);
-      font-weight: 500;
-      padding: 4px 8px;
-      line-height: 1;
-      flex-shrink: 0;
-    }
-
-    [data-slot="content"] {
-      display: flex;
-      align-items: center;
-      gap: 1ch;
-    }
-
-    [data-slot="text"] {
-      color: var(--color-text-strong);
-      line-height: 1.4;
-
-      @media (max-width: 30.625rem) {
-        display: none;
-      }
-    }
-  }
-
   [data-slot="hero-copy"] {
     img {
       margin-bottom: 24px;
@@ -662,10 +631,6 @@ body {
         fill: var(--color-text-strong);
       }
 
-      [data-bar][data-kind="promo"] {
-        fill: color-mix(in srgb, var(--bar-go) 50%, transparent);
-      }
-
       [data-val] {
         fill: var(--color-text-strong);
         font-size: 13px;
diff --git a/packages/console/app/src/routes/go/index.tsx b/packages/console/app/src/routes/go/index.tsx
index 67ae58ae88..71102c7227 100644
--- a/packages/console/app/src/routes/go/index.tsx
+++ b/packages/console/app/src/routes/go/index.tsx
@@ -27,8 +27,6 @@ const models = [
   { name: "GLM-5", provider: "DeepInfra, Fireworks AI, Z.ai" },
   { name: "Kimi K2.5", provider: "Moonshot AI" },
   { name: "Kimi K2.6", provider: "Moonshot AI" },
-  { name: "MiMo-V2-Pro", provider: "Xiaomi MiMo" },
-  { name: "MiMo-V2-Omni", provider: "Xiaomi MiMo" },
   { name: "MiMo-V2.5-Pro", provider: "Xiaomi MiMo" },
   { name: "MiMo-V2.5", provider: "Xiaomi MiMo" },
   { name: "Qwen3.5 Plus", provider: "Alibaba Cloud Model Studio" },
@@ -63,7 +61,7 @@ function LimitsGraph(props: { href: string }) {
   const free = 200
   const graph = [
     { id: "glm-5.1", name: "GLM-5.1", req: 880, d: "100ms" },
-    { id: "kimi-k2.6", name: "Kimi K2.6 (3x usage)", req: 3450, baseReq: 1150, d: "150ms" },
+    { id: "kimi-k2.6", name: "Kimi K2.6", req: 1150, d: "150ms" },
     { id: "mimo-v2.5-pro", name: "MiMo-V2.5-Pro", req: 1290, d: "150ms" },
     { id: "qwen3.6-plus", name: "Qwen3.6 Plus", req: 3300, d: "280ms" },
     { id: "minimax-m2.7", name: "MiniMax M2.7", req: 3400, d: "300ms" },
@@ -157,24 +155,12 @@ function LimitsGraph(props: { href: string }) {
                   <rect
                     x={left}
                     y={gy(i()) - bh / 2}
-                    width={Math.max(0, x(ratio(m.baseReq ?? m.req)) - left)}
+                    width={Math.max(0, x(ratio(m.req)) - left)}
                     height={bh}
                     data-bar
                     data-kind="go"
                     data-model={m.id}
-                    data-segment={m.baseReq ? "base" : undefined}
                   />
-                  {m.baseReq && (
-                    <rect
-                      x={x(ratio(m.baseReq)) + 2}
-                      y={gy(i()) - bh / 2}
-                      width={Math.max(0, x(ratio(m.req)) - x(ratio(m.baseReq)) - 2)}
-                      height={bh}
-                      data-bar
-                      data-kind="promo"
-                      data-model={m.id}
-                    />
-                  )}
                 </g>
               )}
             </For>
@@ -264,12 +250,6 @@ export default function Home() {
 
         <div data-component="content">
           <section data-component="hero">
-            <div data-component="desktop-app-banner">
-              <span data-slot="badge">{i18n.t("home.banner.badge")}</span>
-              <div data-slot="content">
-                <span data-slot="text">{i18n.t("go.banner.text")}</span>
-              </div>
-            </div>
             <div data-slot="hero-copy">
               <img data-slot="zen logo light" src={goLogoLight} alt="" />
               <img data-slot="zen logo dark" src={goLogoDark} alt="" />
diff --git a/packages/console/app/src/routes/honeycomb/webhook.ts b/packages/console/app/src/routes/honeycomb/webhook.ts
new file mode 100644
index 0000000000..367a93aeb0
--- /dev/null
+++ b/packages/console/app/src/routes/honeycomb/webhook.ts
@@ -0,0 +1,85 @@
+import type { APIEvent } from "@solidjs/start/server"
+import { z } from "zod"
+import { Resource } from "@opencode-ai/console-resource"
+import { safeEqual } from "@opencode-ai/console-core/util/crypto.js"
+
+const DISCORD_ALERT_ROLE_ID = "1501447160175136838"
+
+const basePayload = z.object({
+  name: z.string().optional(),
+  status: z.string().optional(),
+  isTest: z.boolean().optional(),
+  url: z.string(),
+})
+
+const groups = z.object({ group: z.object({ key: z.string(), value: z.string() }).array() }).array()
+
+const honeycombWebhookPayload = z.discriminatedUnion("type", [
+  basePayload.extend({
+    type: z.literal("model_http_errors"),
+    groups,
+  }),
+  basePayload.extend({
+    type: z.literal("provider_http_errors"),
+    groups,
+  }),
+  basePayload.extend({
+    type: z.literal("custom"),
+  }),
+])
+
+const postDiscordMessage = async (payload: z.infer<typeof honeycombWebhookPayload>) => {
+  const group =
+    payload.type === "model_http_errors" ? "model" : payload.type === "provider_http_errors" ? "provider" : undefined
+  const names = payload.type === "custom" ? [] : payload.groups.flatMap((item) => item.group.map((g) => g.value))
+
+  const content = [
+    `[**${payload.isTest ? "[TEST] " : ""}${payload.name ?? "Honeycomb alert"}**](${payload.url})`,
+    group && names.length > 0 ? `Affected ${group}s:` : undefined,
+    ...names.map((name) => `- ${name}`),
+    "",
+    `<@&${DISCORD_ALERT_ROLE_ID}>`,
+  ]
+    .filter((line) => line !== undefined)
+    .join("\n")
+
+  return fetch(Resource.DISCORD_INCIDENT_WEBHOOK_URL.value, {
+    method: "POST",
+    headers: { "Content-Type": "application/json" },
+    body: JSON.stringify({
+      content,
+      allowed_mentions: { roles: [DISCORD_ALERT_ROLE_ID] },
+      flags: 4,
+    }),
+  })
+}
+
+export async function POST(input: APIEvent) {
+  const token = input.request.headers.get("X-Honeycomb-Webhook-Token")
+  if (!safeEqual(token ?? "", Resource.HoneycombWebhookSecret.value)) {
+    console.debug("Invalid Honeycomb webhook token")
+    return Response.json({ message: "invalid token" }, { status: 401 })
+  }
+
+  const body = await input.request.json()
+  console.log(body, JSON.stringify(body, null, 2))
+
+  const parsed = honeycombWebhookPayload.safeParse(body)
+
+  if (!parsed.success) {
+    console.error(parsed.error)
+    return Response.json({ message: "invalid payload" }, { status: 400 })
+  }
+
+  if (parsed.data.status !== "TRIGGERED") {
+    console.debug("Skipping resolved alert Honeycomb webhook")
+    return Response.json({ message: "ignored" }, { status: 200 })
+  }
+
+  const response = await postDiscordMessage(parsed.data)
+  if (!response.ok) {
+    return Response.json({ message: "discord webhook failed" }, { status: 502 })
+  }
+
+  return Response.json({ message: "sent" }, { status: 200 })
+}
diff --git a/packages/console/app/src/routes/workspace/[id]/go/lite-section.tsx b/packages/console/app/src/routes/workspace/[id]/go/lite-section.tsx
index 0df181ae16..eba52b0e17 100644
--- a/packages/console/app/src/routes/workspace/[id]/go/lite-section.tsx
+++ b/packages/console/app/src/routes/workspace/[id]/go/lite-section.tsx
@@ -289,8 +289,6 @@ export function LiteSection() {
             <li>Kimi K2.6</li>
             <li>GLM-5</li>
             <li>GLM-5.1</li>
-            <li>MiMo-V2-Pro</li>
-            <li>MiMo-V2-Omni</li>
             <li>MiMo-V2.5-Pro</li>
             <li>MiMo-V2.5</li>
             <li>MiniMax M2.5</li>
diff --git a/packages/console/app/src/routes/workspace/[id]/model-section.tsx b/packages/console/app/src/routes/workspace/[id]/model-section.tsx
index b9cdf3bc3a..35ea2cf878 100644
--- a/packages/console/app/src/routes/workspace/[id]/model-section.tsx
+++ b/packages/console/app/src/routes/workspace/[id]/model-section.tsx
@@ -45,6 +45,7 @@ const getModelsInfo = query(async (workspaceID: string) => {
       all: Object.entries(ZenData.list("full").models)
         .filter(([id, _model]) => !["claude-3-5-haiku"].includes(id))
         .filter(([id, _model]) => !id.startsWith("alpha-"))
+        .filter(([id, _model]) => !id.endsWith(":global"))
         .sort(([idA, modelA], [idB, modelB]) => {
           const priority = ["big-pickle", "minimax", "grok", "claude", "gpt", "gemini"]
           const getPriority = (id: string) => {
diff --git a/packages/console/app/src/routes/zen/util/error.ts b/packages/console/app/src/routes/zen/util/error.ts
index b2a1d30d03..216b6564e7 100644
--- a/packages/console/app/src/routes/zen/util/error.ts
+++ b/packages/console/app/src/routes/zen/util/error.ts
@@ -13,4 +13,13 @@ class LimitError extends Error {
 }
 export class RateLimitError extends LimitError {}
 export class FreeUsageLimitError extends LimitError {}
-export class SubscriptionUsageLimitError extends LimitError {}
+
+class SubscriptionUsageLimitError extends LimitError {
+  workspace: string
+  constructor(message: string, workspace: string, retryAfter?: number) {
+    super(message, retryAfter)
+    this.workspace = workspace
+  }
+}
+export class GoUsageLimitError extends SubscriptionUsageLimitError {}
+export class BlackUsageLimitError extends SubscriptionUsageLimitError {}
diff --git a/packages/console/app/src/routes/zen/util/handler.ts b/packages/console/app/src/routes/zen/util/handler.ts
index 7f36246ee5..c12129ff1d 100644
--- a/packages/console/app/src/routes/zen/util/handler.ts
+++ b/packages/console/app/src/routes/zen/util/handler.ts
@@ -23,7 +23,8 @@ import {
   ModelError,
   RateLimitError,
   FreeUsageLimitError,
-  SubscriptionUsageLimitError,
+  GoUsageLimitError,
+  BlackUsageLimitError,
 } from "./error"
 import {
   buildCostChunk,
@@ -116,7 +117,7 @@ export async function handler(
     const trialProviders = await trialLimiter?.check()
     const rateLimiter = modelInfo.allowAnonymous
       ? createIpRateLimiter(modelInfo.id, modelInfo.rateLimit, ip, input.request)
-      : createKeyRateLimiter(modelInfo.id, zenApiKey, input.request)
+      : createKeyRateLimiter(modelInfo.id, modelInfo.rateLimit, zenApiKey, input.request)
     await rateLimiter?.check()
     const stickyTracker = createStickyTracker(modelInfo.stickyProvider, sessionId)
     const stickyProvider = await stickyTracker?.get()
@@ -395,7 +396,8 @@ export async function handler(
     if (
       error instanceof RateLimitError ||
       error instanceof FreeUsageLimitError ||
-      error instanceof SubscriptionUsageLimitError
+      error instanceof GoUsageLimitError ||
+      error instanceof BlackUsageLimitError
     ) {
       const headers = new Headers()
       if (error.retryAfter) {
@@ -404,7 +406,14 @@ export async function handler(
       return new Response(
         JSON.stringify({
           type: "error",
-          error: { type: error.constructor.name, message: error.message },
+          error: {
+            type: error.constructor.name,
+            message: error.message,
+          },
+          metadata:
+            error instanceof GoUsageLimitError || error instanceof BlackUsageLimitError
+              ? { workspace: error.workspace }
+              : {},
         }),
         { status: 429, headers },
       )
@@ -693,10 +702,11 @@ export async function handler(
             timeUpdated: sub.timeFixedUpdated,
           })
           if (result.status === "rate-limited")
-            throw new SubscriptionUsageLimitError(
+            throw new BlackUsageLimitError(
               t("zen.api.error.subscriptionQuotaExceeded", {
                 retryIn: formatRetryTime(result.resetInSec),
               }),
+              authInfo.workspaceID,
               result.resetInSec,
             )
         }
@@ -711,10 +721,11 @@ export async function handler(
             timeUpdated: sub.timeRollingUpdated,
           })
           if (result.status === "rate-limited")
-            throw new SubscriptionUsageLimitError(
+            throw new BlackUsageLimitError(
               t("zen.api.error.subscriptionQuotaExceeded", {
                 retryIn: formatRetryTime(result.resetInSec),
               }),
+              authInfo.workspaceID,
               result.resetInSec,
             )
         }
@@ -739,8 +750,9 @@ export async function handler(
             timeUpdated: sub.timeWeeklyUpdated,
           })
           if (result.status === "rate-limited")
-            throw new SubscriptionUsageLimitError(
+            throw new GoUsageLimitError(
               t("zen.api.error.subscriptionQuotaExceededUseFreeModels"),
+              authInfo.workspaceID,
               result.resetInSec,
             )
         }
@@ -754,8 +766,9 @@ export async function handler(
             timeSubscribed: sub.timeCreated,
           })
           if (result.status === "rate-limited")
-            throw new SubscriptionUsageLimitError(
+            throw new GoUsageLimitError(
               t("zen.api.error.subscriptionQuotaExceededUseFreeModels"),
+              authInfo.workspaceID,
               result.resetInSec,
             )
         }
@@ -769,8 +782,9 @@ export async function handler(
             timeUpdated: sub.timeRollingUpdated,
           })
           if (result.status === "rate-limited")
-            throw new SubscriptionUsageLimitError(
+            throw new GoUsageLimitError(
               t("zen.api.error.subscriptionQuotaExceededUseFreeModels"),
+              authInfo.workspaceID,
               result.resetInSec,
             )
         }
diff --git a/packages/console/app/src/routes/zen/util/keyRateLimiter.ts b/packages/console/app/src/routes/zen/util/keyRateLimiter.ts
index e3e0fb18f2..0bf495f7db 100644
--- a/packages/console/app/src/routes/zen/util/keyRateLimiter.ts
+++ b/packages/console/app/src/routes/zen/util/keyRateLimiter.ts
@@ -4,11 +4,16 @@ import { RateLimitError } from "./error"
 import { i18n } from "~/i18n"
 import { localeFromRequest } from "~/lib/language"
 
-export function createRateLimiter(modelId: string, zenApiKey: string | undefined, request: Request) {
+export function createRateLimiter(
+  modelId: string,
+  rateLimit: number | undefined,
+  zenApiKey: string | undefined,
+  request: Request,
+) {
   if (!zenApiKey) return
   const dict = i18n(localeFromRequest(request))
 
-  const LIMIT = 100
+  const LIMIT = rateLimit ?? 300
   const yyyyMMddHHmm = new Date(Date.now())
     .toISOString()
     .replace(/[^0-9]/g, "")
diff --git a/packages/console/app/src/routes/zen/v1/models.ts b/packages/console/app/src/routes/zen/v1/models.ts
index 794f85029a..68c3cac694 100644
--- a/packages/console/app/src/routes/zen/v1/models.ts
+++ b/packages/console/app/src/routes/zen/v1/models.ts
@@ -28,7 +28,9 @@ export async function GET(input: APIEvent) {
     )
   })()
 
-  const models = Object.keys(ZenData.list("full").models).filter((id) => !disabledModels.includes(id))
+  const models = Object.keys(ZenData.list("full").models)
+    .filter((id) => !id.endsWith(":global"))
+    .filter((id) => !disabledModels.includes(id))
 
   return buildModelsResponse(models)
 }
diff --git a/packages/console/core/package.json b/packages/console/core/package.json
index bdfc576fb9..4ca29eb4c7 100644
--- a/packages/console/core/package.json
+++ b/packages/console/core/package.json
@@ -1,7 +1,7 @@
 {
   "$schema": "https://json.schemastore.org/package.json",
   "name": "@opencode-ai/console-core",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "private": true,
   "type": "module",
   "license": "MIT",
diff --git a/packages/console/core/src/util/crypto.ts b/packages/console/core/src/util/crypto.ts
new file mode 100644
index 0000000000..46f53ae391
--- /dev/null
+++ b/packages/console/core/src/util/crypto.ts
@@ -0,0 +1,8 @@
+import { timingSafeEqual } from "node:crypto"
+
+export function safeEqual(a: string, b: string): boolean {
+  const encoder = new TextEncoder()
+  const aBytes = encoder.encode(a)
+  const bBytes = encoder.encode(b)
+  return aBytes.length === bBytes.length && timingSafeEqual(aBytes, bBytes)
+}
diff --git a/packages/console/core/sst-env.d.ts b/packages/console/core/sst-env.d.ts
index 288e73d0cb..9680a53aab 100644
--- a/packages/console/core/sst-env.d.ts
+++ b/packages/console/core/sst-env.d.ts
@@ -35,6 +35,10 @@ declare module "sst" {
       "type": "sst.cloudflare.SolidStart"
       "url": string
     }
+    "DISCORD_INCIDENT_WEBHOOK_URL": {
+      "type": "sst.sst.Secret"
+      "value": string
+    }
     "DISCORD_SUPPORT_BOT_TOKEN": {
       "type": "sst.sst.Secret"
       "value": string
@@ -87,6 +91,10 @@ declare module "sst" {
       "type": "sst.sst.Secret"
       "value": string
     }
+    "HoneycombWebhookSecret": {
+      "type": "random.index/randomPassword.RandomPassword"
+      "value": string
+    }
     "R2AccessKey": {
       "type": "sst.sst.Secret"
       "value": string
diff --git a/packages/console/function/package.json b/packages/console/function/package.json
index dc56d8bc29..7e1d77d7dc 100644
--- a/packages/console/function/package.json
+++ b/packages/console/function/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@opencode-ai/console-function",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "$schema": "https://json.schemastore.org/package.json",
   "private": true,
   "type": "module",
diff --git a/packages/console/function/sst-env.d.ts b/packages/console/function/sst-env.d.ts
index 288e73d0cb..9680a53aab 100644
--- a/packages/console/function/sst-env.d.ts
+++ b/packages/console/function/sst-env.d.ts
@@ -35,6 +35,10 @@ declare module "sst" {
       "type": "sst.cloudflare.SolidStart"
       "url": string
     }
+    "DISCORD_INCIDENT_WEBHOOK_URL": {
+      "type": "sst.sst.Secret"
+      "value": string
+    }
     "DISCORD_SUPPORT_BOT_TOKEN": {
       "type": "sst.sst.Secret"
       "value": string
@@ -87,6 +91,10 @@ declare module "sst" {
       "type": "sst.sst.Secret"
       "value": string
     }
+    "HoneycombWebhookSecret": {
+      "type": "random.index/randomPassword.RandomPassword"
+      "value": string
+    }
     "R2AccessKey": {
       "type": "sst.sst.Secret"
       "value": string
diff --git a/packages/console/mail/package.json b/packages/console/mail/package.json
index 1600bb877d..34ddd073f0 100644
--- a/packages/console/mail/package.json
+++ b/packages/console/mail/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@opencode-ai/console-mail",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "dependencies": {
     "@jsx-email/all": "2.2.3",
     "@jsx-email/cli": "1.4.3",
diff --git a/packages/console/resource/sst-env.d.ts b/packages/console/resource/sst-env.d.ts
index 288e73d0cb..9680a53aab 100644
--- a/packages/console/resource/sst-env.d.ts
+++ b/packages/console/resource/sst-env.d.ts
@@ -35,6 +35,10 @@ declare module "sst" {
       "type": "sst.cloudflare.SolidStart"
       "url": string
     }
+    "DISCORD_INCIDENT_WEBHOOK_URL": {
+      "type": "sst.sst.Secret"
+      "value": string
+    }
     "DISCORD_SUPPORT_BOT_TOKEN": {
       "type": "sst.sst.Secret"
       "value": string
@@ -87,6 +91,10 @@ declare module "sst" {
       "type": "sst.sst.Secret"
       "value": string
     }
+    "HoneycombWebhookSecret": {
+      "type": "random.index/randomPassword.RandomPassword"
+      "value": string
+    }
     "R2AccessKey": {
       "type": "sst.sst.Secret"
       "value": string
diff --git a/packages/core/package.json b/packages/core/package.json
index 4c880779e2..01d8c8a4ed 100644
--- a/packages/core/package.json
+++ b/packages/core/package.json
@@ -1,6 +1,6 @@
 {
   "$schema": "https://json.schemastore.org/package.json",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "name": "@opencode-ai/core",
   "type": "module",
   "license": "MIT",
diff --git a/packages/desktop/README.md b/packages/desktop/README.md
index ebaf488223..6dd9a202ad 100644
--- a/packages/desktop/README.md
+++ b/packages/desktop/README.md
@@ -1,32 +1,19 @@
 # OpenCode Desktop
 
-Native OpenCode desktop app, built with Tauri v2.
+The OpenCode Desktop app, built with Electron.
 
 ## Development
 
-From the repo root:
-
 ```bash
 bun install
-bun run --cwd packages/desktop tauri dev
-```
-
-This starts the Vite dev server on http://localhost:1420 and opens the native window.
-
-If you only want the web dev server (no native shell):
-
-```bash
-bun run --cwd packages/desktop dev
+bun dev
 ```
 
 ## Build
 
-To create a production `dist/` and build the native app bundle:
+Run the `build` script to build the app's JS assets, then `package` to
+bundle the assets as an application. The resulting app will be in `dist/`.
 
 ```bash
-bun run --cwd packages/desktop tauri build
+bun run build && bun run package
 ```
-
-## Prerequisites
-
-Running the desktop app requires additional Tauri dependencies (Rust toolchain, platform-specific libraries). See the [Tauri prerequisites](https://v2.tauri.app/start/prerequisites/) for setup instructions.
diff --git a/packages/desktop/electron.vite.config.ts b/packages/desktop/electron.vite.config.ts
index a352e03fdd..52aa699ff6 100644
--- a/packages/desktop/electron.vite.config.ts
+++ b/packages/desktop/electron.vite.config.ts
@@ -37,7 +37,7 @@ export default defineConfig({
     },
     build: {
       rollupOptions: {
-        input: { index: "src/main/index.ts" },
+        input: { index: "src/main/index.ts", sidecar: "src/main/sidecar.ts" },
       },
       externalizeDeps: { include: [nodePtyPkg] },
     },
diff --git a/packages/desktop/package.json b/packages/desktop/package.json
index cbc20b9061..49e35c5db8 100644
--- a/packages/desktop/package.json
+++ b/packages/desktop/package.json
@@ -1,7 +1,7 @@
 {
   "name": "@opencode-ai/desktop",
   "private": true,
-  "version": "1.14.39",
+  "version": "1.14.41",
   "type": "module",
   "license": "MIT",
   "homepage": "https://opencode.ai",
@@ -63,6 +63,14 @@
     "@lydell/node-pty-linux-arm64": "1.2.0-beta.10",
     "@lydell/node-pty-linux-x64": "1.2.0-beta.10",
     "@lydell/node-pty-win32-arm64": "1.2.0-beta.10",
-    "@lydell/node-pty-win32-x64": "1.2.0-beta.10"
+    "@lydell/node-pty-win32-x64": "1.2.0-beta.10",
+    "@parcel/watcher-darwin-arm64": "2.5.1",
+    "@parcel/watcher-darwin-x64": "2.5.1",
+    "@parcel/watcher-linux-arm64-glibc": "2.5.1",
+    "@parcel/watcher-linux-arm64-musl": "2.5.1",
+    "@parcel/watcher-linux-x64-glibc": "2.5.1",
+    "@parcel/watcher-linux-x64-musl": "2.5.1",
+    "@parcel/watcher-win32-arm64": "2.5.1",
+    "@parcel/watcher-win32-x64": "2.5.1"
   }
 }
diff --git a/packages/desktop/src/main/apps.ts b/packages/desktop/src/main/apps.ts
index 174da94a5d..bf25417b83 100644
--- a/packages/desktop/src/main/apps.ts
+++ b/packages/desktop/src/main/apps.ts
@@ -1,14 +1,22 @@
-import { execFileSync } from "node:child_process"
-import { existsSync, readFileSync, readdirSync } from "node:fs"
+import { execFile, execFileSync } from "node:child_process"
+import { access, readFile, readdir } from "node:fs/promises"
 import { dirname, extname, join } from "node:path"
+import util from "node:util"
 
-export function checkAppExists(appName: string): boolean {
+const execFilePromise = util.promisify(execFile)
+
+const exists = (path: string) =>
+  access(path)
+    .then(() => true)
+    .catch(() => false)
+
+export function checkAppExists(appName: string) {
   if (process.platform === "win32") return true
   if (process.platform === "linux") return true
   return checkMacosApp(appName)
 }
 
-export function resolveAppPath(appName: string): string | null {
+export function resolveAppPath(appName: string) {
   if (process.platform !== "win32") return appName
   return resolveWindowsAppPath(appName)
 }
@@ -32,26 +40,25 @@ export function wslPath(path: string, mode: "windows" | "linux" | null): string
   }
 }
 
-function checkMacosApp(appName: string) {
+async function checkMacosApp(appName: string) {
   const locations = [`/Applications/${appName}.app`, `/System/Applications/${appName}.app`]
 
   const home = process.env.HOME
   if (home) locations.push(`${home}/Applications/${appName}.app`)
 
-  if (locations.some((location) => existsSync(location))) return true
-
-  try {
-    execFileSync("which", [appName])
-    return true
-  } catch {
-    return false
+  for (const location of locations) {
+    if (await exists(location)) return true
   }
+
+  return execFilePromise("which", [appName])
+    .then(() => true)
+    .catch(() => false)
 }
 
-function resolveWindowsAppPath(appName: string): string | null {
+async function resolveWindowsAppPath(appName: string): Promise<string | null> {
   let output: string
   try {
-    output = execFileSync("where", [appName]).toString()
+    output = execFilePromise("where", [appName]).toString()
   } catch {
     return null
   }
@@ -66,8 +73,8 @@ function resolveWindowsAppPath(appName: string): string | null {
   const exe = paths.find((path) => hasExt(path, "exe"))
   if (exe) return exe
 
-  const resolveCmd = (path: string) => {
-    const content = readFileSync(path, "utf8")
+  const resolveCmd = async (path: string) => {
+    const content = await readFile(path, "utf8")
     for (const token of content.split('"').map((value: string) => value.trim())) {
       const lower = token.toLowerCase()
       if (!lower.includes(".exe")) continue
@@ -85,10 +92,10 @@ function resolveWindowsAppPath(appName: string): string | null {
             return join(current, part)
           }, base)
 
-        if (existsSync(resolved)) return resolved
+        if (await exists(resolved)) return resolved
       }
 
-      if (existsSync(token)) return token
+      if (await exists(token)) return token
     }
 
     return null
@@ -96,20 +103,20 @@ function resolveWindowsAppPath(appName: string): string | null {
 
   for (const path of paths) {
     if (hasExt(path, "cmd") || hasExt(path, "bat")) {
-      const resolved = resolveCmd(path)
+      const resolved = await resolveCmd(path)
       if (resolved) return resolved
     }
 
     if (!extname(path)) {
       const cmd = `${path}.cmd`
-      if (existsSync(cmd)) {
-        const resolved = resolveCmd(cmd)
+      if (await exists(cmd)) {
+        const resolved = await resolveCmd(cmd)
         if (resolved) return resolved
       }
 
       const bat = `${path}.bat`
-      if (existsSync(bat)) {
-        const resolved = resolveCmd(bat)
+      if (await exists(bat)) {
+        const resolved = await resolveCmd(bat)
         if (resolved) return resolved
       }
     }
@@ -126,7 +133,7 @@ function resolveWindowsAppPath(appName: string): string | null {
       const dirs = [dirname(path), dirname(dirname(path)), dirname(dirname(dirname(path)))]
       for (const dir of dirs) {
         try {
-          for (const entry of readdirSync(dir)) {
+          for (const entry of await readdir(dir)) {
             const candidate = join(dir, entry)
             if (!hasExt(candidate, "exe")) continue
             const stem = entry.replace(/\.exe$/i, "")
diff --git a/packages/desktop/src/main/env.d.ts b/packages/desktop/src/main/env.d.ts
index 1de56e1c90..eee21e48cb 100644
--- a/packages/desktop/src/main/env.d.ts
+++ b/packages/desktop/src/main/env.d.ts
@@ -5,6 +5,7 @@ interface ImportMetaEnv {
 interface ImportMeta {
   readonly env: ImportMetaEnv
 }
+
 declare module "virtual:opencode-server" {
   export namespace Server {
     export const listen: typeof import("../../../opencode/dist/types/src/node").Server.listen
diff --git a/packages/desktop/src/main/index.ts b/packages/desktop/src/main/index.ts
index a1eba8b98d..f75cd719a2 100644
--- a/packages/desktop/src/main/index.ts
+++ b/packages/desktop/src/main/index.ts
@@ -1,9 +1,9 @@
 import { randomUUID } from "node:crypto"
 import { EventEmitter } from "node:events"
-import { existsSync } from "node:fs"
+import { existsSync, mkdirSync, rmSync } from "node:fs"
 import * as http from "node:http"
 import { createServer } from "node:net"
-import { homedir } from "node:os"
+import { homedir, tmpdir } from "node:os"
 import { join } from "node:path"
 import { getCACertificates, setDefaultCACertificates } from "node:tls"
 import type { Event } from "electron"
@@ -30,10 +30,14 @@ const APP_IDS: Record<string, string> = {
   beta: "ai.opencode.desktop.beta",
   prod: "ai.opencode.desktop",
 }
+const TEST_ONBOARDING = process.env.OPENCODE_TEST_ONBOARDING === "1"
 const appId = app.isPackaged ? APP_IDS[CHANNEL] : "ai.opencode.desktop.dev"
+const onboardingTestRoot = setupOnboardingTestEnv()
 app.setName(app.isPackaged ? APP_NAMES[CHANNEL] : "OpenCode Dev")
 app.setAppUserModelId(appId)
-app.setPath("userData", join(app.getPath("appData"), appId))
+app.setPath("userData", onboardingTestRoot ? join(onboardingTestRoot, "desktop") : join(app.getPath("appData"), appId))
+if (onboardingTestRoot) app.setPath("sessionData", join(onboardingTestRoot, "session"))
+const logger = initLogging()
 const { autoUpdater } = pkg
 
 import type { InitStep, ServerReadyData, SqliteMigrationProgress, WslConfig } from "../preload/types"
@@ -43,7 +47,15 @@ import { registerIpcHandlers, sendDeepLinks, sendMenuCommand, sendSqliteMigratio
 import { initLogging } from "./logging"
 import { parseMarkdown } from "./markdown"
 import { createMenu } from "./menu"
-import { getDefaultServerUrl, getWslConfig, setDefaultServerUrl, setWslConfig, spawnLocalServer } from "./server"
+import {
+  getDefaultServerUrl,
+  getWslConfig,
+  preferAppEnv,
+  setDefaultServerUrl,
+  setWslConfig,
+  spawnLocalServer,
+  type SidecarListener,
+} from "./server"
 import {
   createLoadingWindow,
   createMainWindow,
@@ -51,27 +63,41 @@ import {
   setBackgroundColor,
   setDockIcon,
 } from "./windows"
-import { drizzle } from "drizzle-orm/node-sqlite/driver"
-import type { Server } from "virtual:opencode-server"
 import { migrate } from "./migrate"
 
 const initEmitter = new EventEmitter()
 let initStep: InitStep = { phase: "server_waiting" }
 
 let mainWindow: BrowserWindow | null = null
-let server: Server.Listener | null = null
+let server: SidecarListener | null = null
 const loadingComplete = defer<void>()
 
 const pendingDeepLinks: string[] = []
 
 const serverReady = defer<ServerReadyData>()
-const logger = initLogging()
 
 useSystemCertificates()
 
+function setupOnboardingTestEnv() {
+  if (!TEST_ONBOARDING) return
+
+  const root = join(tmpdir(), `opencode-onboarding-${randomUUID()}`)
+  rmSync(root, { recursive: true, force: true })
+  ;["data", "config", "cache", "state", "desktop", "session"].forEach((dir) =>
+    mkdirSync(join(root, dir), { recursive: true }),
+  )
+  process.env.OPENCODE_DB = ":memory:"
+  process.env.XDG_DATA_HOME = join(root, "data")
+  process.env.XDG_CONFIG_HOME = join(root, "config")
+  process.env.XDG_CACHE_HOME = join(root, "cache")
+  process.env.XDG_STATE_HOME = join(root, "state")
+  return root
+}
+
 logger.log("app starting", {
   version: app.getVersion(),
   packaged: app.isPackaged,
+  onboardingTest: Boolean(onboardingTestRoot),
 })
 
 setupApp()
@@ -87,6 +113,8 @@ function setupApp() {
     return
   }
 
+  preferAppEnv(app.getPath("userData"))
+
   app.on("second-instance", (_event: Event, argv: string[]) => {
     const urls = argv.filter((arg: string) => arg.startsWith("opencode://"))
     if (urls.length) {
@@ -103,22 +131,21 @@ function setupApp() {
   })
 
   app.on("before-quit", () => {
-    killSidecar()
+    void killSidecar()
   })
 
   app.on("will-quit", () => {
-    killSidecar()
+    void killSidecar()
   })
 
   for (const signal of ["SIGINT", "SIGTERM"] as const) {
     process.on(signal, () => {
-      killSidecar()
-      app.exit(0)
+      void killSidecar().finally(() => app.exit(0))
     })
   }
 
   void app.whenReady().then(async () => {
-    migrate()
+    if (!TEST_ONBOARDING) migrate()
     app.setAsDefaultProtocolClient("opencode")
     registerRendererProtocol()
     setDockIcon()
@@ -164,7 +191,6 @@ function setInitStep(step: InitStep) {
 
 async function initialize() {
   const needsMigration = !sqliteFileExists()
-  const sqliteDone = needsMigration ? defer<void>() : undefined
   let overlay: BrowserWindow | null = null
 
   const port = await getSidecarPort()
@@ -179,31 +205,26 @@ async function initialize() {
       setInitStep({ phase: "sqlite_waiting" })
       if (overlay) sendSqliteMigrationProgress(overlay, progress)
       if (mainWindow) sendSqliteMigrationProgress(mainWindow, progress)
-      if (progress.type === "Done") sqliteDone?.resolve()
     })
 
-    if (needsMigration) {
-      const { Database, JsonMigration } = await import("virtual:opencode-server")
-      await JsonMigration.run(drizzle({ client: Database.Client().$client }), {
-        progress: (event: { current: number; total: number }) => {
-          const percent = Math.round(event.current / event.total) * 100
-          initEmitter.emit("sqlite", { type: "InProgress", value: percent })
-        },
-      })
-      initEmitter.emit("sqlite", { type: "Done" })
-
-      sqliteDone?.resolve()
-    }
-
-    if (needsMigration) {
-      await sqliteDone?.promise
-    }
-
     logger.log("spawning sidecar", { url })
-    const { listener, health } = await spawnLocalServer(hostname, port, password, () => {
-      ensureLoopbackNoProxy()
-      useEnvProxy()
-    })
+    const { listener, health } = await spawnLocalServer(
+      hostname,
+      port,
+      password,
+      () => {
+        ensureLoopbackNoProxy()
+        useEnvProxy()
+      },
+      {
+        needsMigration,
+        userDataPath: app.getPath("userData"),
+        onSqliteProgress: (progress) => initEmitter.emit("sqlite", progress),
+        onStdout: (message) => logger.log("sidecar stdout", { message }),
+        onStderr: (message) => logger.warn("sidecar stderr", { message }),
+        onExit: (code) => logger.warn("sidecar exited", { code }),
+      },
+    )
     server = listener
     serverReady.resolve({
       url,
@@ -253,9 +274,10 @@ function wireMenu() {
     },
     reload: () => mainWindow?.reload(),
     relaunch: () => {
-      killSidecar()
-      app.relaunch()
-      app.exit(0)
+      void killSidecar().finally(() => {
+        app.relaunch()
+        app.exit(0)
+      })
     },
   })
 }
@@ -284,7 +306,7 @@ registerIpcHandlers({
   getDisplayBackend: async () => null,
   setDisplayBackend: async () => undefined,
   parseMarkdown: async (markdown) => parseMarkdown(markdown),
-  checkAppExists: async (appName) => checkAppExists(appName),
+  checkAppExists: (appName) => checkAppExists(appName),
   wslPath: async (path, mode) => wslPath(path, mode),
   resolveAppPath: async (appName) => resolveAppPath(appName),
   loadingWindowComplete: () => loadingComplete.resolve(),
@@ -294,10 +316,11 @@ registerIpcHandlers({
   setBackgroundColor: (color) => setBackgroundColor(color),
 })
 
-function killSidecar() {
+async function killSidecar() {
   if (!server) return
-  server.stop()
+  const current = server
   server = null
+  await current.stop()
 }
 
 function ensureLoopbackNoProxy() {
@@ -344,6 +367,8 @@ async function getSidecarPort() {
 }
 
 function sqliteFileExists() {
+  if (process.env.OPENCODE_DB === ":memory:") return true
+
   const xdg = process.env.XDG_DATA_HOME
   const base = xdg && xdg.length > 0 ? xdg : join(homedir(), ".local", "share")
   return existsSync(join(base, "opencode", "opencode.db"))
@@ -356,7 +381,7 @@ function setupAutoUpdater() {
   autoUpdater.allowPrerelease = false
   autoUpdater.allowDowngrade = true
   autoUpdater.autoDownload = false
-  autoUpdater.autoInstallOnAppQuit = true
+  autoUpdater.autoInstallOnAppQuit = false
   logger.log("auto updater configured", {
     channel: autoUpdater.channel,
     allowPrerelease: autoUpdater.allowPrerelease,
@@ -418,7 +443,7 @@ async function installUpdate() {
   logger.log("installing downloaded update", {
     version: downloadedUpdateVersion,
   })
-  killSidecar()
+  await killSidecar()
   autoUpdater.quitAndInstall()
 }
 
diff --git a/packages/desktop/src/main/ipc.ts b/packages/desktop/src/main/ipc.ts
index 1c4af0eb60..dbcd4239dc 100644
--- a/packages/desktop/src/main/ipc.ts
+++ b/packages/desktop/src/main/ipc.ts
@@ -19,7 +19,7 @@ const pickerFilters = (ext?: string[]) => {
 }
 
 type Deps = {
-  killSidecar: () => void
+  killSidecar: () => Promise<void> | void
   awaitInitialization: (sendStep: (step: InitStep) => void) => Promise<ServerReadyData>
   getWindowConfig: () => Promise<WindowConfig> | WindowConfig
   consumeInitialDeepLinks: () => Promise<string[]> | string[]
diff --git a/packages/desktop/src/main/logging.ts b/packages/desktop/src/main/logging.ts
index d315b2d344..1f1c5e54e3 100644
--- a/packages/desktop/src/main/logging.ts
+++ b/packages/desktop/src/main/logging.ts
@@ -7,6 +7,7 @@ const TAIL_LINES = 1000
 
 export function initLogging() {
   log.transports.file.maxSize = 5 * 1024 * 1024
+  initConsoleTransport()
   cleanup()
   return log
 }
@@ -38,3 +39,19 @@ function cleanup() {
     }
   }
 }
+
+function initConsoleTransport() {
+  const write = log.transports.console.writeFn.bind(log.transports.console)
+  log.transports.console.writeFn = (options) => {
+    try {
+      write(options)
+    } catch (err) {
+      if (!isBrokenPipe(err)) throw err
+      log.transports.console.level = false
+    }
+  }
+}
+
+function isBrokenPipe(err: unknown) {
+  return typeof err === "object" && err !== null && "code" in err && err.code === "EPIPE"
+}
diff --git a/packages/desktop/src/main/menu.ts b/packages/desktop/src/main/menu.ts
index 0d9a697fa9..2d5a900f39 100644
--- a/packages/desktop/src/main/menu.ts
+++ b/packages/desktop/src/main/menu.ts
@@ -23,6 +23,11 @@ export function createMenu(deps: Deps) {
           enabled: UPDATER_ENABLED,
           click: () => deps.checkForUpdates(),
         },
+        {
+          label: "Settings",
+          accelerator: "Cmd+,",
+          click: () => deps.trigger("settings.open"),
+        },
         {
           label: "Reload Webview",
           click: () => deps.reload(),
diff --git a/packages/desktop/src/main/server.ts b/packages/desktop/src/main/server.ts
index fab09eb1b1..635a93578a 100644
--- a/packages/desktop/src/main/server.ts
+++ b/packages/desktop/src/main/server.ts
@@ -1,12 +1,37 @@
-import { app } from "electron"
+import { dirname, join } from "node:path"
+import { fileURLToPath } from "node:url"
+import { app, utilityProcess } from "electron"
+import type { Details } from "electron"
 import { DEFAULT_SERVER_URL_KEY, WSL_ENABLED_KEY } from "./constants"
-import { getUserShell, loadShellEnv } from "./shell-env"
+import { getUserShell, loadShellEnv, mergeShellEnv } from "./shell-env"
 import { getStore } from "./store"
+import type { SqliteMigrationProgress } from "../preload/types"
 
 export type WslConfig = { enabled: boolean }
 
 export type HealthCheck = { wait: Promise<void> }
 
+type SidecarMessage =
+  | { type: "sqlite"; progress: SqliteMigrationProgress }
+  | { type: "ready" }
+  | { type: "stopped" }
+  | { type: "error"; error: { message: string; stack?: string } }
+
+export type SidecarListener = { stop: () => Promise<void> }
+
+const SIDECAR_SERVICE_NAME = "opencode server"
+const SIDECAR_START_STALL_TIMEOUT = 60_000
+const SIDECAR_STOP_TIMEOUT = 6_000
+
+type SpawnLocalServerOptions = {
+  needsMigration: boolean
+  userDataPath: string
+  onSqliteProgress?: (progress: SqliteMigrationProgress) => void
+  onStdout?: (message: string) => void
+  onStderr?: (message: string) => void
+  onExit?: (code: number) => void
+}
+
 export function getDefaultServerUrl(): string | null {
   const value = getStore().get(DEFAULT_SERVER_URL_KEY)
   return typeof value === "string" ? value : null
@@ -30,49 +55,155 @@ export function setWslConfig(config: WslConfig) {
   getStore().set(WSL_ENABLED_KEY, config.enabled)
 }
 
-export async function spawnLocalServer(hostname: string, port: number, password: string, configureEnv?: () => void) {
-  prepareServerEnv(password)
+export function preferAppEnv(userDataPath: string) {
+  const shell = process.platform === "win32" ? null : getUserShell()
+  Object.assign(
+    process.env,
+    mergeShellEnv(shell ? loadShellEnv(shell) : null, {
+      ...process.env,
+      OPENCODE_EXPERIMENTAL_ICON_DISCOVERY: "true",
+      OPENCODE_EXPERIMENTAL_FILEWATCHER: "true",
+      OPENCODE_CLIENT: "desktop",
+      XDG_STATE_HOME: process.env.XDG_STATE_HOME ?? userDataPath,
+    }),
+  )
+}
+
+export async function spawnLocalServer(
+  hostname: string,
+  port: number,
+  password: string,
+  configureEnv: () => void,
+  options: SpawnLocalServerOptions,
+) {
   configureEnv?.()
-  const { Log, Server } = await import("virtual:opencode-server")
-  await Log.init({ level: "WARN" })
-  const listener = await Server.listen({
-    port,
-    hostname,
-    username: "opencode",
-    password,
-    cors: ["oc://renderer"],
+  const sidecar = join(dirname(fileURLToPath(import.meta.url)), "sidecar.js")
+  const child = utilityProcess.fork(sidecar, [], {
+    cwd: process.cwd(),
+    env: createSidecarEnv(),
+    serviceName: SIDECAR_SERVICE_NAME,
+    stdio: "pipe",
+  })
+  let exited = false
+  const exit = defer<number>()
+
+  const onProcessGone = (_event: unknown, details: Details) => {
+    if (details.type !== "Utility" || details.name !== SIDECAR_SERVICE_NAME) return
+    options.onStderr?.(`utility process gone reason=${details.reason} exitCode=${details.exitCode}`)
+  }
+
+  app.on("child-process-gone", onProcessGone)
+  child.once("exit", (code) => {
+    exited = true
+    app.off("child-process-gone", onProcessGone)
+    options.onExit?.(code)
+    exit.resolve(code)
+  })
+  child.on("error", (error) => options.onStderr?.(`utility process error: ${serializeError(error).message}`))
+
+  child.stdout?.on("data", (chunk: Buffer) => options.onStdout?.(chunk.toString("utf8").trimEnd()))
+  child.stderr?.on("data", (chunk: Buffer) => options.onStderr?.(chunk.toString("utf8").trimEnd()))
+
+  await new Promise<void>((resolve, reject) => {
+    let done = false
+    let timeout: NodeJS.Timeout
+
+    const fail = (error: Error) => {
+      if (done) return
+      done = true
+      cleanup()
+      reject(error)
+    }
+
+    const refreshTimeout = () => {
+      clearTimeout(timeout)
+      timeout = setTimeout(() => {
+        fail(new Error(`Sidecar did not become ready within ${SIDECAR_START_STALL_TIMEOUT}ms: ${sidecar}`))
+      }, SIDECAR_START_STALL_TIMEOUT)
+    }
+
+    const onMessage = (message: SidecarMessage) => {
+      if (message.type === "sqlite") {
+        refreshTimeout()
+        options.onSqliteProgress?.(message.progress)
+        return
+      }
+      if (message.type === "ready") {
+        if (done) return
+        done = true
+        cleanup()
+        resolve()
+        return
+      }
+      if (message.type === "error") {
+        fail(Object.assign(new Error(message.error.message), { stack: message.error.stack }))
+      }
+    }
+    const onExit = (code: number) => {
+      fail(new Error(`Sidecar exited before ready with code ${code}`))
+    }
+    const cleanup = () => {
+      clearTimeout(timeout)
+      child.off("message", onMessage)
+      child.off("exit", onExit)
+    }
+
+    child.on("message", onMessage)
+    child.on("exit", onExit)
+    refreshTimeout()
+    child.postMessage({
+      type: "start",
+      hostname,
+      port,
+      password,
+      userDataPath: options.userDataPath,
+      needsMigration: options.needsMigration,
+    })
+  }).catch((error) => {
+    if (!exited) child.kill()
+    throw error
   })
 
   const wait = (async () => {
     const url = `http://${hostname}:${port}`
+    let healthy = false
+    const gone = exit.promise.then((code) => {
+      if (healthy) return
+      throw new Error(`Sidecar exited before health check passed with code ${code}`)
+    })
 
     const ready = async () => {
       while (true) {
         await new Promise((resolve) => setTimeout(resolve, 100))
-        if (await checkHealth(url, password)) return
+        if (await checkHealth(url, password)) {
+          healthy = true
+          return
+        }
       }
     }
 
-    await ready()
+    await Promise.race([ready(), gone])
   })()
 
-  return { listener, health: { wait } }
-}
+  let stopping: Promise<void> | undefined
 
-function prepareServerEnv(password: string) {
-  const shell = process.platform === "win32" ? null : getUserShell()
-  const shellEnv = shell ? (loadShellEnv(shell) ?? {}) : {}
-  const env = {
-    ...process.env,
-    ...shellEnv,
-    OPENCODE_EXPERIMENTAL_ICON_DISCOVERY: "true",
-    OPENCODE_EXPERIMENTAL_FILEWATCHER: "true",
-    OPENCODE_CLIENT: "desktop",
-    OPENCODE_SERVER_USERNAME: "opencode",
-    OPENCODE_SERVER_PASSWORD: password,
-    XDG_STATE_HOME: app.getPath("userData"),
+  return {
+    listener: {
+      stop: () => {
+        if (stopping) return stopping
+        if (exited) return Promise.resolve()
+        child.postMessage({ type: "stop" })
+        stopping = Promise.race([
+          exit.promise.then(() => undefined),
+          delay(SIDECAR_STOP_TIMEOUT).then(() => {
+            if (!exited) child.kill()
+          }),
+        ])
+        return stopping
+      },
+    },
+    health: { wait },
   }
-  Object.assign(process.env, env)
 }
 
 export async function checkHealth(url: string, password?: string | null): Promise<boolean> {
@@ -100,3 +231,31 @@ export async function checkHealth(url: string, password?: string | null): Promis
     return false
   }
 }
+
+function createSidecarEnv(): Record<string, string> {
+  const env = Object.fromEntries(
+    Object.entries(process.env).flatMap(([key, value]) => (value === undefined ? [] : [[key, String(value)]])),
+  )
+  delete env.DEBUG
+  if (process.platform === "linux") delete env.LD_PRELOAD
+  return env
+}
+
+function delay(ms: number) {
+  return new Promise<void>((resolve) => setTimeout(resolve, ms))
+}
+
+function serializeError(error: unknown) {
+  if (error instanceof Error) return { message: error.message, stack: error.stack }
+  return { message: String(error) }
+}
+
+function defer<T>() {
+  let resolve!: (value: T) => void
+  let reject!: (error: Error) => void
+  const promise = new Promise<T>((res, rej) => {
+    resolve = res
+    reject = rej
+  })
+  return { promise, resolve, reject }
+}
diff --git a/packages/desktop/src/main/sidecar.ts b/packages/desktop/src/main/sidecar.ts
new file mode 100644
index 0000000000..e7d652b6e1
--- /dev/null
+++ b/packages/desktop/src/main/sidecar.ts
@@ -0,0 +1,178 @@
+import { drizzle } from "drizzle-orm/node-sqlite/driver"
+import * as http from "node:http"
+import * as tls from "node:tls"
+
+type NodeHttpWithEnvProxy = typeof http & {
+  setGlobalProxyFromEnv: () => void
+}
+
+type NodeTlsWithSystemCertificates = typeof tls & {
+  getCACertificates: (type: "default" | "system") => string[]
+  setDefaultCACertificates: (certificates: string[]) => void
+}
+
+type StartCommand = {
+  type: "start"
+  hostname: string
+  port: number
+  password: string
+  userDataPath: string
+  needsMigration: boolean
+}
+
+type StopCommand = { type: "stop" }
+type SidecarCommand = StartCommand | StopCommand
+
+type SidecarMessage =
+  | { type: "sqlite"; progress: { type: "InProgress"; value: number } | { type: "Done" } }
+  | { type: "ready" }
+  | { type: "stopped" }
+  | { type: "error"; error: { message: string; stack?: string } }
+
+type ParentPort = {
+  postMessage(message: SidecarMessage): void
+  on(event: "message", listener: (event: { data: unknown }) => void): void
+}
+
+type Listener = {
+  stop(close?: boolean): void | Promise<void>
+}
+
+const parentPort = getParentPort()
+let listener: Listener | undefined
+
+parentPort.on("message", (event) => {
+  const command = parseCommand(event.data)
+  if (!command) return
+  if (command.type === "stop") {
+    void stop()
+    return
+  }
+  void start(command)
+})
+
+async function start(command: StartCommand) {
+  try {
+    prepareSidecarEnv(command.password, command.userDataPath)
+    ensureLoopbackNoProxy()
+    useSystemCertificates()
+    useEnvProxy()
+    const { Database, JsonMigration, Log, Server } = await import("virtual:opencode-server")
+    await Log.init({ level: "WARN" })
+
+    if (command.needsMigration) {
+      await JsonMigration.run(drizzle({ client: Database.Client().$client }), {
+        progress: (event: { current: number; total: number }) => {
+          parentPort.postMessage({
+            type: "sqlite",
+            progress: {
+              type: "InProgress",
+              value: event.total === 0 ? 100 : Math.round((event.current / event.total) * 100),
+            },
+          })
+        },
+      })
+      parentPort.postMessage({ type: "sqlite", progress: { type: "Done" } })
+    }
+
+    listener = await Server.listen({
+      port: command.port,
+      hostname: command.hostname,
+      username: "opencode",
+      password: command.password,
+      cors: ["oc://renderer"],
+    })
+    parentPort.postMessage({ type: "ready" })
+  } catch (error) {
+    parentPort.postMessage({ type: "error", error: serializeError(error) })
+    setImmediate(() => process.exit(1))
+  }
+}
+
+async function stop() {
+  try {
+    await listener?.stop()
+  } finally {
+    listener = undefined
+    parentPort.postMessage({ type: "stopped" })
+    setImmediate(() => process.exit(0))
+  }
+}
+
+function prepareSidecarEnv(password: string, userDataPath: string) {
+  Object.assign(process.env, {
+    OPENCODE_SERVER_USERNAME: "opencode",
+    OPENCODE_SERVER_PASSWORD: password,
+    XDG_STATE_HOME: process.env.XDG_STATE_HOME ?? userDataPath,
+  })
+}
+
+function ensureLoopbackNoProxy() {
+  const loopback = ["127.0.0.1", "localhost", "::1"]
+  const upsert = (key: string) => {
+    const items = (process.env[key] ?? "")
+      .split(",")
+      .map((value: string) => value.trim())
+      .filter((value: string) => Boolean(value))
+
+    for (const host of loopback) {
+      if (items.some((value: string) => value.toLowerCase() === host)) continue
+      items.push(host)
+    }
+
+    process.env[key] = items.join(",")
+  }
+
+  upsert("NO_PROXY")
+  upsert("no_proxy")
+}
+
+function useSystemCertificates() {
+  try {
+    const nodeTls = tls as NodeTlsWithSystemCertificates
+    nodeTls.setDefaultCACertificates([
+      ...new Set([...nodeTls.getCACertificates("default"), ...nodeTls.getCACertificates("system")]),
+    ])
+  } catch (error) {
+    console.warn("failed to load system certificates", error)
+  }
+}
+
+function useEnvProxy() {
+  try {
+    ;(http as NodeHttpWithEnvProxy).setGlobalProxyFromEnv()
+  } catch (error) {
+    console.warn("failed to load proxy environment", error)
+  }
+}
+
+function parseCommand(value: unknown): SidecarCommand | undefined {
+  if (!value || typeof value !== "object") return
+  const command = value as Partial<StartCommand | StopCommand>
+  if (command.type === "stop") return { type: "stop" }
+  if (command.type !== "start") return
+  if (typeof command.hostname !== "string") return
+  if (typeof command.port !== "number") return
+  if (typeof command.password !== "string") return
+  if (typeof command.userDataPath !== "string") return
+  if (typeof command.needsMigration !== "boolean") return
+  return {
+    type: "start",
+    hostname: command.hostname,
+    port: command.port,
+    password: command.password,
+    userDataPath: command.userDataPath,
+    needsMigration: command.needsMigration,
+  }
+}
+
+function serializeError(error: unknown) {
+  if (error instanceof Error) return { message: error.message, stack: error.stack }
+  return { message: String(error) }
+}
+
+function getParentPort() {
+  const port = process.parentPort as ParentPort | undefined
+  if (!port) throw new Error("Sidecar parent port unavailable")
+  return port
+}
diff --git a/packages/desktop/src/main/store.ts b/packages/desktop/src/main/store.ts
index 7b3bd7c660..a591f878de 100644
--- a/packages/desktop/src/main/store.ts
+++ b/packages/desktop/src/main/store.ts
@@ -1,4 +1,5 @@
 import Store from "electron-store"
+import { app } from "electron"
 
 import { SETTINGS_STORE } from "./constants"
 
@@ -11,7 +12,12 @@ const cache = new Map<string, Store>()
 export function getStore(name = SETTINGS_STORE) {
   const cached = cache.get(name)
   if (cached) return cached
-  const next = new Store({ name, fileExtension: "", accessPropertiesByDotNotation: false })
+  const next = new Store({
+    name,
+    cwd: app.getPath("userData"),
+    fileExtension: "",
+    accessPropertiesByDotNotation: false,
+  })
   cache.set(name, next)
   return next
 }
diff --git a/packages/desktop/src/main/windows.ts b/packages/desktop/src/main/windows.ts
index 387e793b0e..41abfc784d 100644
--- a/packages/desktop/src/main/windows.ts
+++ b/packages/desktop/src/main/windows.ts
@@ -8,6 +8,7 @@ const root = dirname(fileURLToPath(import.meta.url))
 const rendererRoot = join(root, "../renderer")
 const rendererProtocol = "oc"
 const rendererHost = "renderer"
+const clipboardWritePermission = "clipboard-sanitized-write"
 
 protocol.registerSchemesAsPrivileged([
   {
@@ -107,6 +108,8 @@ export function createMainWindow() {
     },
   })
 
+  allowClipboardWrite(win)
+
   win.webContents.session.webRequest.onBeforeSendHeaders((details, callback) => {
     const { requestHeaders } = details
     upsertKeyValue(requestHeaders, "Access-Control-Allow-Origin", ["*"])
@@ -157,6 +160,8 @@ export function createLoadingWindow() {
     },
   })
 
+  allowClipboardWrite(win)
+
   loadWindow(win, "loading.html")
 
   return win
@@ -191,6 +196,31 @@ function loadWindow(win: BrowserWindow, html: string) {
 
   void win.loadURL(`${rendererProtocol}://${rendererHost}/${html}`)
 }
+
+function allowClipboardWrite(win: BrowserWindow) {
+  win.webContents.session.setPermissionRequestHandler((webContents, permission, callback, details) => {
+    callback(
+      permission === clipboardWritePermission &&
+        isTrustedRendererUrl(details.requestingUrl) &&
+        webContents.id === win.webContents.id,
+    )
+  })
+  win.webContents.session.setPermissionCheckHandler((webContents, permission, requestingOrigin, details) => {
+    if (permission !== clipboardWritePermission) return false
+    if (webContents && webContents.id !== win.webContents.id) return false
+    return isTrustedRendererUrl(details.requestingUrl) || isTrustedRendererUrl(requestingOrigin)
+  })
+}
+
+function isTrustedRendererUrl(value?: string) {
+  if (!value || !URL.canParse(value)) return false
+  const url = new URL(value)
+  if (url.protocol === `${rendererProtocol}:` && url.host === rendererHost) return true
+  const devUrl = process.env.ELECTRON_RENDERER_URL
+  if (!devUrl || !URL.canParse(devUrl)) return false
+  return url.origin === new URL(devUrl).origin
+}
+
 function wireZoom(win: BrowserWindow) {
   win.webContents.setZoomFactor(1)
   win.webContents.on("zoom-changed", () => {
diff --git a/packages/desktop/src/renderer/index.tsx b/packages/desktop/src/renderer/index.tsx
index 97c7ed23a2..f9114c7550 100644
--- a/packages/desktop/src/renderer/index.tsx
+++ b/packages/desktop/src/renderer/index.tsx
@@ -43,7 +43,11 @@ if (import.meta.env.VITE_SENTRY_DSN) {
     integrations: (integrations) => {
       return integrations.filter(
         (i) =>
-          i.name !== "Breadcrumbs" && !(import.meta.env.OPENCODE_CHANNEL === "prod" && i.name === "GlobalHandlers"),
+          i.name !== "Breadcrumbs" &&
+          !(
+            import.meta.env.OPENCODE_CHANNEL === "prod" &&
+            (i.name === "GlobalHandlers" || i.name === "BrowserApiErrors")
+          ),
       )
     },
   })
diff --git a/packages/enterprise/package.json b/packages/enterprise/package.json
index 49509aa075..beccdb6991 100644
--- a/packages/enterprise/package.json
+++ b/packages/enterprise/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@opencode-ai/enterprise",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "private": true,
   "type": "module",
   "license": "MIT",
diff --git a/packages/enterprise/sst-env.d.ts b/packages/enterprise/sst-env.d.ts
index 288e73d0cb..9680a53aab 100644
--- a/packages/enterprise/sst-env.d.ts
+++ b/packages/enterprise/sst-env.d.ts
@@ -35,6 +35,10 @@ declare module "sst" {
       "type": "sst.cloudflare.SolidStart"
       "url": string
     }
+    "DISCORD_INCIDENT_WEBHOOK_URL": {
+      "type": "sst.sst.Secret"
+      "value": string
+    }
     "DISCORD_SUPPORT_BOT_TOKEN": {
       "type": "sst.sst.Secret"
       "value": string
@@ -87,6 +91,10 @@ declare module "sst" {
       "type": "sst.sst.Secret"
       "value": string
     }
+    "HoneycombWebhookSecret": {
+      "type": "random.index/randomPassword.RandomPassword"
+      "value": string
+    }
     "R2AccessKey": {
       "type": "sst.sst.Secret"
       "value": string
diff --git a/packages/extensions/zed/extension.toml b/packages/extensions/zed/extension.toml
index 8102023128..8b4850c885 100644
--- a/packages/extensions/zed/extension.toml
+++ b/packages/extensions/zed/extension.toml
@@ -1,7 +1,7 @@
 id = "opencode"
 name = "OpenCode"
 description = "The open source coding agent."
-version = "1.14.39"
+version = "1.14.41"
 schema_version = 1
 authors = ["Anomaly"]
 repository = "https://github.com/anomalyco/opencode"
@@ -11,26 +11,26 @@ name = "OpenCode"
 icon = "./icons/opencode.svg"
 
 [agent_servers.opencode.targets.darwin-aarch64]
-archive = "https://github.com/anomalyco/opencode/releases/download/v1.14.39/opencode-darwin-arm64.zip"
+archive = "https://github.com/anomalyco/opencode/releases/download/v1.14.41/opencode-darwin-arm64.zip"
 cmd = "./opencode"
 args = ["acp"]
 
 [agent_servers.opencode.targets.darwin-x86_64]
-archive = "https://github.com/anomalyco/opencode/releases/download/v1.14.39/opencode-darwin-x64.zip"
+archive = "https://github.com/anomalyco/opencode/releases/download/v1.14.41/opencode-darwin-x64.zip"
 cmd = "./opencode"
 args = ["acp"]
 
 [agent_servers.opencode.targets.linux-aarch64]
-archive = "https://github.com/anomalyco/opencode/releases/download/v1.14.39/opencode-linux-arm64.tar.gz"
+archive = "https://github.com/anomalyco/opencode/releases/download/v1.14.41/opencode-linux-arm64.tar.gz"
 cmd = "./opencode"
 args = ["acp"]
 
 [agent_servers.opencode.targets.linux-x86_64]
-archive = "https://github.com/anomalyco/opencode/releases/download/v1.14.39/opencode-linux-x64.tar.gz"
+archive = "https://github.com/anomalyco/opencode/releases/download/v1.14.41/opencode-linux-x64.tar.gz"
 cmd = "./opencode"
 args = ["acp"]
 
 [agent_servers.opencode.targets.windows-x86_64]
-archive = "https://github.com/anomalyco/opencode/releases/download/v1.14.39/opencode-windows-x64.zip"
+archive = "https://github.com/anomalyco/opencode/releases/download/v1.14.41/opencode-windows-x64.zip"
 cmd = "./opencode.exe"
 args = ["acp"]
diff --git a/packages/function/package.json b/packages/function/package.json
index 84219c5510..70812ab10a 100644
--- a/packages/function/package.json
+++ b/packages/function/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@opencode-ai/function",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "$schema": "https://json.schemastore.org/package.json",
   "private": true,
   "type": "module",
diff --git a/packages/function/sst-env.d.ts b/packages/function/sst-env.d.ts
index 288e73d0cb..9680a53aab 100644
--- a/packages/function/sst-env.d.ts
+++ b/packages/function/sst-env.d.ts
@@ -35,6 +35,10 @@ declare module "sst" {
       "type": "sst.cloudflare.SolidStart"
       "url": string
     }
+    "DISCORD_INCIDENT_WEBHOOK_URL": {
+      "type": "sst.sst.Secret"
+      "value": string
+    }
     "DISCORD_SUPPORT_BOT_TOKEN": {
       "type": "sst.sst.Secret"
       "value": string
@@ -87,6 +91,10 @@ declare module "sst" {
       "type": "sst.sst.Secret"
       "value": string
     }
+    "HoneycombWebhookSecret": {
+      "type": "random.index/randomPassword.RandomPassword"
+      "value": string
+    }
     "R2AccessKey": {
       "type": "sst.sst.Secret"
       "value": string
diff --git a/packages/opencode/package.json b/packages/opencode/package.json
index 459c7d7a85..1951dfb481 100644
--- a/packages/opencode/package.json
+++ b/packages/opencode/package.json
@@ -1,6 +1,6 @@
 {
   "$schema": "https://json.schemastore.org/package.json",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "name": "@browser-use/browsercode-core",
   "type": "module",
   "license": "MIT",
@@ -80,7 +80,7 @@
   "dependencies": {
     "@actions/core": "1.11.1",
     "@actions/github": "6.0.1",
-    "@agentclientprotocol/sdk": "0.16.1",
+    "@agentclientprotocol/sdk": "0.21.0",
     "@ai-sdk/alibaba": "1.0.17",
     "@ai-sdk/amazon-bedrock": "4.0.96",
     "@ai-sdk/anthropic": "3.0.71",
diff --git a/packages/opencode/specs/effect/errors.md b/packages/opencode/specs/effect/errors.md
new file mode 100644
index 0000000000..746e658693
--- /dev/null
+++ b/packages/opencode/specs/effect/errors.md
@@ -0,0 +1,329 @@
+# Typed error migration
+
+Plan for moving `packages/opencode` from temporary defect/`NamedError`
+compatibility toward typed Effect service errors and explicit HTTP error
+contracts.
+
+## Goal
+
+- Expected service failures live on the Effect error channel.
+- Service interfaces expose those failures in their return types.
+- Domain errors are authored with Effect Schema so they are reusable by services,
+  tests, HTTP routes, tools, and OpenAPI generation.
+- HTTP status codes and wire compatibility are handled at the HTTP boundary, not
+  inside service modules.
+- `Effect.die`, `throw`, `catchDefect`, and global cause inspection are reserved
+  for defects, compatibility bridges, or final fallback behavior.
+
+## Current State
+
+- Many migrated services use Effect internally, but expected failures are still a
+  mix of `NamedError.create(...)`, `namedSchemaError(...)`, `class extends Error`,
+  `throw`, and `Effect.die(...)`.
+- Some services already use `Schema.TaggedErrorClass`, for example `Account`,
+  `Auth`, `Permission`, `Question`, `Installation`, and parts of
+  `Workspace`.
+- Legacy Hono error handling recognizes `NamedError`, `Session.BusyError`, and a
+  few name-based cases, then emits the legacy `{ name, data }` JSON body.
+- Effect `HttpApi` only knows how to encode errors that are declared on the
+  endpoint, group, or middleware. Undeclared expected errors become defects and
+  eventually fall through to generic HTTP handling.
+- The temporary HttpApi error middleware catches defect-wrapped legacy errors to
+  preserve runtime behavior, but it is intentionally a bridge rather than the
+  final model.
+
+## End State
+
+Service modules own domain failures.
+
+```ts
+export class SessionBusyError extends Schema.TaggedErrorClass<SessionBusyError>()("SessionBusyError", {
+  sessionID: SessionID,
+  message: Schema.String,
+}) {}
+
+export type Error = Storage.Error | SessionBusyError
+
+export interface Interface {
+  readonly get: (id: SessionID) => Effect.Effect<Info, Error>
+}
+```
+
+HTTP modules own transport mapping.
+
+```ts
+const get = Effect.fn("SessionHttpApi.get")(function* (ctx: { params: { sessionID: SessionID } }) {
+  return yield* session
+    .get(ctx.params.sessionID)
+    .pipe(
+      Effect.catchTag("StorageNotFoundError", () => new SessionNotFoundHttpError({ sessionID: ctx.params.sessionID })),
+    )
+})
+```
+
+HTTP-visible error schemas carry their own response status through Effect
+HttpApi's `httpApiStatus` annotation. Prefer `HttpApiSchema.status(...)`, or the
+equivalent declaration annotation, instead of maintaining a parallel status map.
+
+```ts
+export class SessionNotFoundHttpError extends Schema.TaggedErrorClass<SessionNotFoundHttpError>()(
+  "SessionNotFoundHttpError",
+  {
+    sessionID: SessionID,
+    message: Schema.String,
+  },
+  { httpApiStatus: 404 },
+) {}
+```
+
+Endpoint definitions still declare which HTTP-visible error schemas can be
+emitted. The status annotation is only used if the error is part of the endpoint,
+group, or middleware error schema and the handler fails with that error on the
+typed error channel.
+
+```ts
+HttpApiEndpoint.get("get", SessionPaths.get, {
+  success: Session.Info,
+  error: [SessionNotFoundHttpError, SessionBusyHttpError],
+})
+```
+
+The service error and HTTP error may be the same class when the wire shape is a
+deliberate public contract. They should be different classes when the service
+error contains internals, low-level causes, retry hints, or anything that should
+not be exposed to API clients.
+
+## Rules
+
+- Use `Schema.TaggedErrorClass` for new expected domain errors.
+- Include `cause: Schema.optional(Schema.Defect)` only when preserving an
+  underlying unknown failure is useful for logs or callers.
+- Export a domain-level error union from each service module, for example
+  `export type Error = NotFoundError | BusyError | Storage.Error`.
+- Put expected errors in service method signatures, for example
+  `Effect.Effect<Result, Service.Error, R>`.
+- Use `yield* new DomainError(...)` for direct early failures inside
+  `Effect.gen` / `Effect.fn`.
+- Use `Effect.try({ try, catch })`, `Effect.mapError`, or `Effect.catchTag` to
+  convert external exceptions into domain errors.
+- Use `HttpApiSchema.status(...)` or `{ httpApiStatus: code }` on HTTP-visible
+  error schemas so Effect `HttpApiBuilder` and OpenAPI generation get the status
+  from the schema itself.
+- Do not use `Effect.die(...)` for user, IO, validation, missing-resource, auth,
+  provider, worktree, or busy-state failures.
+- Do not use `catchDefect` to recover expected domain errors. If recovery is
+  needed, the upstream effect should fail with a typed error instead.
+- Do not make service modules import `HttpApiError`, `HttpServerResponse`, HTTP
+  status codes, or route-specific error schemas.
+- Keep raw `HttpRouter` routes free to use `HttpServerRespondable` when that is
+  the right transport abstraction, but prefer declared `HttpApi` errors for
+  normal JSON API endpoints.
+
+## HTTP Boundary Shape
+
+Create an HttpApi-local error module, likely
+`src/server/routes/instance/httpapi/errors.ts`.
+
+That module should provide:
+
+- Legacy-compatible public schemas for `{ name, data }` error bodies that must
+  remain SDK-compatible during the Hono migration.
+- Small constructors or mapping helpers for common API errors such as not found,
+  bad request, conflict, and unknown internal errors.
+- Route-group-specific adapters only when they encode domain-specific public
+  data.
+- A single place to document which public error shape is legacy-compatible and
+  which shape is new Effect-native API surface.
+
+Avoid one giant `unknown -> status` mapper. Prefer small, explicit mappers close
+to the handler or route group.
+
+```ts
+const mapSessionError = <A, E, R>(effect: Effect.Effect<A, E, R>) =>
+  effect.pipe(
+    Effect.catchTag("StorageNotFoundError", (error) => new SessionNotFoundHttpError({ message: error.message })),
+    Effect.catchTag("SessionBusyError", (error) => new SessionBusyHttpError({ message: error.message })),
+  )
+```
+
+Use built-in `HttpApiError.BadRequest`, `HttpApiError.NotFound`, and related
+types only when their generated response body and SDK surface are intentionally
+acceptable. Use a custom schema-backed error when clients need the legacy
+`{ name, data }` body or a domain-specific error payload.
+
+## Migration Phases
+
+### 1. Stabilize The Bridge
+
+Keep the temporary HttpApi error middleware only as a compatibility bridge while
+typed errors are introduced.
+
+- Add tests that prove the bridge catches legacy `NamedError` defects.
+- Add tests that prove declared HttpApi errors still use the declared endpoint
+  contract.
+- Stop returning stack traces in unknown HTTP `500` responses; log the full
+  `Cause.pretty(cause)` server-side instead.
+- Add a comment or TODO that names this plan and states the bridge must shrink
+  as route groups migrate.
+
+### 2. Define The Shared HTTP Error Helpers
+
+Add the `httpapi/errors.ts` module before converting route groups.
+
+- Define a legacy `{ name, data }` body helper for SDK-compatible errors.
+- Define `UnknownError` for generic internal failures with a safe public message.
+- Define `BadRequestError` and `NotFoundError` equivalents only if the actual
+  wire body must match the legacy Hono SDK surface.
+- Put the HTTP status on the public schema with `HttpApiSchema.status(...)` or
+  `{ httpApiStatus: code }`; do not keep a separate name-to-status table.
+- Keep conversion helpers pure and small. They should not inspect `Cause` or
+  accept `unknown` unless they are final fallback helpers.
+
+### 3. Convert One Vertical Slice
+
+Start with session read routes because they already have local `mapNotFound`
+logic and are heavily covered by existing HttpApi tests.
+
+- Convert `Session.BusyError` from a plain `Error` to a typed service error, or
+  add a typed wrapper while preserving the old constructor until callers are
+  migrated.
+- Replace `catchDefect` in `httpapi/handlers/session.ts` with typed error
+  mapping.
+- Add endpoint error schemas for the affected session endpoints.
+- Prove behavior with focused tests in `test/server/httpapi-session.test.ts`.
+- Remove the migrated cases from the global compatibility middleware.
+
+### 4. Convert Legacy NamedError Domains
+
+Move legacy `NamedError.create(...)` services to Effect Schema-backed errors in
+small domain PRs.
+
+Priority order:
+
+1. `storage/storage.ts` and `storage/db.ts` not-found errors.
+2. `worktree/index.ts` `Worktree*` errors.
+3. `provider/auth.ts` validation failures and `provider/provider.ts` model-not-found errors.
+4. `mcp/index.ts`, `skill/index.ts`, `lsp/client.ts`, and `ide/index.ts` service errors.
+5. Config and CLI-only errors after HTTP-facing domains are stable.
+
+For each domain:
+
+- Replace `NamedError.create(...)` with `Schema.TaggedErrorClass` when the error
+  is primarily a service error.
+- Keep or add a separate HTTP error schema when the legacy `{ name, data }` wire
+  shape must remain stable.
+- Update service interface return types to include the new error union.
+- Replace `throw new X(...)` inside `Effect.fn` with `yield* new X(...)`.
+- Replace async exceptions with `Effect.try({ catch })` or explicit `mapError`.
+- Add service-level tests that assert the error tag and data, not just the HTTP
+  status.
+
+### 5. Declare HttpApi Errors Group By Group
+
+For each HttpApi group:
+
+- Inventory every service call and the typed errors it can return.
+- Add only the public error schemas that endpoint can actually emit.
+- Map service errors to HTTP errors in the handler file.
+- Keep built-in `HttpApiError` only for generic request/validation failures where
+  the generated contract is accepted.
+- Update `httpapi/public.ts` compatibility transforms only when the generated
+  spec cannot represent the desired source shape directly.
+- Regenerate the SDK after OpenAPI-visible changes and verify the diff is
+  intentional.
+
+Suggested route order:
+
+1. `session` not-found and busy-state reads.
+2. `experimental` worktree mutations.
+3. `provider` auth and model selection errors.
+4. `mcp` OAuth and connection errors.
+5. Remaining route groups as Hono deletion work progresses.
+
+### 6. Remove Defect Recovery
+
+After enough route groups declare their expected errors:
+
+- Delete `catchDefect` recovery for domain errors.
+- Delete name-prefix checks such as `error.name.startsWith("Worktree")` from
+  HTTP middleware.
+- Delete `NamedError` branches from the Effect HttpApi compatibility middleware
+  once no Effect route depends on them.
+- Leave one final unknown-defect fallback that logs server-side and returns a
+  safe generic `500` body.
+
+## Inventory Checklist
+
+Use this checklist when touching a service or route group.
+
+- [ ] Does the service interface expose every expected failure in the Effect
+      error type?
+- [ ] Are user-caused, provider-caused, IO, auth, missing-resource, and busy-state
+      failures modeled as typed errors instead of defects?
+- [ ] Does the service avoid importing HTTP status, `HttpApiError`, or response
+      classes?
+- [ ] Does the handler map each service error into a declared endpoint error?
+- [ ] Does the endpoint `error` field include every public error the handler can
+      emit?
+- [ ] Does OpenAPI/SDK output either stay byte-identical or have an explicitly
+      reviewed diff?
+- [ ] Do tests cover both service-level error typing and HTTP-level status/body?
+- [ ] Did the PR remove any now-unneeded case from the temporary compatibility
+      middleware?
+
+## Testing Requirements
+
+For service conversions:
+
+- Test the service method directly with `testEffect(...)`.
+- Assert on `_tag` or class identity and the structured fields.
+- Avoid testing by string-matching `Cause.pretty(...)`.
+
+For HttpApi conversions:
+
+- Add or update the focused `test/server/httpapi-*.test.ts` file.
+- Assert status code, content type, and exact JSON body for declared public
+  errors.
+- Add a regression test that the temporary middleware is no longer needed for the
+  migrated route.
+- Keep bridge/parity tests aligned with legacy Hono behavior until Hono is
+  deleted or the SDK contract intentionally changes.
+
+## Verification Commands
+
+Run from `packages/opencode` unless noted otherwise.
+
+```bash
+bun run prettier --write <changed files>
+bunx oxlint <changed files>
+bun typecheck
+bun run test -- test/server/httpapi-session.test.ts
+```
+
+Run SDK generation from the repo root when schemas or OpenAPI-visible errors
+change.
+
+```bash
+./packages/sdk/js/script/build.ts
+```
+
+## Open Questions
+
+- Should legacy V1 routes keep `{ name, data }` forever while V2 routes expose a
+  more Effect-native tagged error body?
+- Should storage not-found remain generic, or should callers map it to
+  domain-specific not-found errors before crossing service boundaries?
+- Should `namedSchemaError(...)` stay as a long-term public-wire helper, or only
+  as a migration bridge for old `NamedError` contracts?
+- Which SDK version boundary lets us stop remapping built-in Effect HttpApi error
+  schemas in `httpapi/public.ts`?
+
+## Success Criteria
+
+- New service code no longer uses `die` for expected failures.
+- A route reviewer can read an endpoint definition and see every public error it
+  can return.
+- The temporary HttpApi error middleware shrinks over time instead of gaining new
+  name-based cases.
+- Service tests prove domain error types without going through HTTP.
+- HTTP tests prove status/body contracts without relying on defect recovery.
diff --git a/packages/opencode/src/acp/agent.ts b/packages/opencode/src/acp/agent.ts
index d66c1b2583..ad930680d1 100644
--- a/packages/opencode/src/acp/agent.ts
+++ b/packages/opencode/src/acp/agent.ts
@@ -5,6 +5,8 @@ import {
   type AuthenticateRequest,
   type AuthMethod,
   type CancelNotification,
+  type CloseSessionRequest,
+  type CloseSessionResponse,
   type ForkSessionRequest,
   type ForkSessionResponse,
   type InitializeRequest,
@@ -565,6 +567,7 @@ export class Agent implements ACPAgent {
           image: true,
         },
         sessionCapabilities: {
+          close: {},
           fork: {},
           list: {},
           resume: {},
@@ -627,6 +630,9 @@ export class Agent implements ACPAgent {
       // Store ACP session state
       await this.sessionManager.load(sessionId, params.cwd, params.mcpServers, model)
 
+      const messages = await this.loadSessionMessages(directory, sessionId)
+      this.restoreSessionStateFromMessages(sessionId, messages)
+
       log.info("load_session", { sessionId, mcpServers: params.mcpServers.length })
 
       const result = await this.loadSessionMode({
@@ -635,39 +641,6 @@ export class Agent implements ACPAgent {
         sessionId,
       })
 
-      // Replay session history
-      const messages = await this.sdk.session
-        .messages(
-          {
-            sessionID: sessionId,
-            directory,
-          },
-          { throwOnError: true },
-        )
-        .then((x) => x.data)
-        .catch((err) => {
-          log.error("unexpected error when fetching message", { error: err })
-          return undefined
-        })
-
-      const lastUser = messages?.findLast((m) => m.info.role === "user")?.info
-      if (lastUser?.role === "user") {
-        result.models.currentModelId = `${lastUser.model.providerID}/${lastUser.model.modelID}`
-        this.sessionManager.setModel(sessionId, {
-          providerID: ProviderID.make(lastUser.model.providerID),
-          modelID: ModelID.make(lastUser.model.modelID),
-        })
-        if (result.modes?.availableModes.some((m) => m.id === lastUser.agent)) {
-          result.modes.currentModeId = lastUser.agent
-          this.sessionManager.setMode(sessionId, lastUser.agent)
-        }
-        result.configOptions = buildConfigOptions({
-          currentModelId: result.models.currentModelId,
-          availableModels: result.models.availableModels,
-          modes: result.modes,
-        })
-      }
-
       for (const msg of messages ?? []) {
         log.debug("replay message", msg)
         await this.processMessage(msg)
@@ -756,6 +729,9 @@ export class Agent implements ACPAgent {
       const sessionId = forked.id
       await this.sessionManager.load(sessionId, directory, mcpServers, model)
 
+      const messages = await this.loadSessionMessages(directory, sessionId)
+      this.restoreSessionStateFromMessages(sessionId, messages)
+
       log.info("fork_session", { sessionId, mcpServers: mcpServers.length })
 
       const mode = await this.loadSessionMode({
@@ -764,20 +740,6 @@ export class Agent implements ACPAgent {
         sessionId,
       })
 
-      const messages = await this.sdk.session
-        .messages(
-          {
-            sessionID: sessionId,
-            directory,
-          },
-          { throwOnError: true },
-        )
-        .then((x) => x.data)
-        .catch((err) => {
-          log.error("unexpected error when fetching message", { error: err })
-          return undefined
-        })
-
       for (const msg of messages ?? []) {
         log.debug("replay message", msg)
         await this.processMessage(msg)
@@ -797,7 +759,7 @@ export class Agent implements ACPAgent {
     }
   }
 
-  async unstable_resumeSession(params: ResumeSessionRequest): Promise<ResumeSessionResponse> {
+  async resumeSession(params: ResumeSessionRequest): Promise<ResumeSessionResponse> {
     const directory = params.cwd
     const sessionId = params.sessionId
     const mcpServers = params.mcpServers ?? []
@@ -806,6 +768,9 @@ export class Agent implements ACPAgent {
       const model = await defaultModel(this.config, directory)
       await this.sessionManager.load(sessionId, directory, mcpServers, model)
 
+      const messages = await this.loadSessionMessages(directory, sessionId, 20)
+      this.restoreSessionStateFromMessages(sessionId, messages)
+
       log.info("resume_session", { sessionId, mcpServers: mcpServers.length })
 
       const result = await this.loadSessionMode({
@@ -828,6 +793,27 @@ export class Agent implements ACPAgent {
     }
   }
 
+  async closeSession(params: CloseSessionRequest): Promise<CloseSessionResponse> {
+    const session = this.sessionManager.remove(params.sessionId)
+    if (!session) return {}
+
+    await this.sdk.session
+      .abort(
+        {
+          sessionID: params.sessionId,
+          directory: session.cwd,
+        },
+        { throwOnError: true },
+      )
+      .catch((error) => {
+        log.error("failed to abort session while closing ACP session", { error, sessionID: params.sessionId })
+      })
+
+    this.permissionQueues.delete(params.sessionId)
+    log.info("close_session", { sessionId: params.sessionId })
+    return {}
+  }
+
   private async processMessage(message: SessionMessageResponse) {
     log.debug("process message", message)
     if (message.info.role !== "assistant" && message.info.role !== "user") return
@@ -1159,23 +1145,26 @@ export class Agent implements ACPAgent {
     sessionId: string,
   ): Promise<{ availableModes: ModeOption[]; currentModeId?: string }> {
     const availableModes = await this.loadAvailableModes(directory)
-    const currentModeId =
-      this.sessionManager.get(sessionId).modeId ||
-      (await (async () => {
-        if (!availableModes.length) return undefined
-        const defaultAgentName = await AppRuntime.runPromise(AgentModule.Service.use((svc) => svc.defaultAgent()))
-        const resolvedModeId = availableModes.find((mode) => mode.name === defaultAgentName)?.id ?? availableModes[0].id
-        this.sessionManager.setMode(sessionId, resolvedModeId)
-        return resolvedModeId
-      })())
+    const storedModeId = this.sessionManager.get(sessionId).modeId
+    if (storedModeId && availableModes.some((mode) => mode.id === storedModeId)) {
+      return { availableModes, currentModeId: storedModeId }
+    }
+
+    const currentModeId = await (async () => {
+      if (!availableModes.length) return undefined
+      const defaultAgentName = await AppRuntime.runPromise(AgentModule.Service.use((svc) => svc.defaultAgent()))
+      const resolvedModeId = availableModes.find((mode) => mode.name === defaultAgentName)?.id ?? availableModes[0].id
+      this.sessionManager.setMode(sessionId, resolvedModeId)
+      return resolvedModeId
+    })()
 
     return { availableModes, currentModeId }
   }
 
   private async loadSessionMode(params: LoadSessionRequest) {
     const directory = params.cwd
-    const model = await defaultModel(this.config, directory)
     const sessionId = params.sessionId
+    const model = this.sessionManager.get(sessionId).model ?? (await defaultModel(this.config, directory))
 
     const providers = await this.sdk.config.providers({ directory }).then((x) => x.data!.providers)
     const entries = sortProvidersByName(providers)
@@ -1184,7 +1173,7 @@ export class Agent implements ACPAgent {
     if (currentVariant && !availableVariants.includes(currentVariant)) {
       this.sessionManager.setVariant(sessionId, undefined)
     }
-    const availableModels = buildAvailableModels(entries, { includeVariants: true })
+    const availableModels = buildAvailableModels(entries)
     const modeState = await this.resolveModeState(directory, sessionId)
     const currentModeId = modeState.currentModeId
     const modes = currentModeId
@@ -1267,13 +1256,15 @@ export class Agent implements ACPAgent {
     return {
       sessionId,
       models: {
-        currentModelId: formatModelIdWithVariant(model, currentVariant, availableVariants, true),
+        currentModelId: formatModelIdWithVariant(model, currentVariant, availableVariants, false),
         availableModels,
       },
       modes,
       configOptions: buildConfigOptions({
-        currentModelId: formatModelIdWithVariant(model, currentVariant, availableVariants, true),
+        currentModelId: formatModelIdWithVariant(model, currentVariant, availableVariants, false),
         availableModels,
+        currentVariant,
+        availableVariants,
         modes,
       }),
       _meta: buildVariantMeta({
@@ -1296,6 +1287,24 @@ export class Agent implements ACPAgent {
 
     const entries = sortProvidersByName(providers)
     const availableVariants = modelVariantsFromProviders(entries, selection.model)
+    const modeState = await this.resolveModeState(session.cwd, session.id)
+    const modes = modeState.currentModeId
+      ? { availableModes: modeState.availableModes, currentModeId: modeState.currentModeId }
+      : undefined
+
+    await this.connection.sessionUpdate({
+      sessionId: session.id,
+      update: {
+        sessionUpdate: "config_option_update",
+        configOptions: buildConfigOptions({
+          currentModelId: formatModelIdWithVariant(selection.model, selection.variant, availableVariants, false),
+          availableModels: buildAvailableModels(entries),
+          currentVariant: selection.variant,
+          availableVariants,
+          modes,
+        }),
+      },
+    })
 
     return {
       _meta: buildVariantMeta({
@@ -1327,6 +1336,14 @@ export class Agent implements ACPAgent {
       const selection = parseModelSelection(params.value, providers)
       this.sessionManager.setModel(session.id, selection.model)
       this.sessionManager.setVariant(session.id, selection.variant)
+    } else if (params.configId === "effort") {
+      if (typeof params.value !== "string") throw RequestError.invalidParams("effort value must be a string")
+      const current = session.model ?? (await defaultModel(this.config, session.cwd))
+      const availableVariants = modelVariantsFromProviders(entries, current)
+      if (!availableVariants.includes(params.value)) {
+        throw RequestError.invalidParams(JSON.stringify({ error: `Effort not found: ${params.value}` }))
+      }
+      this.sessionManager.setVariant(session.id, params.value)
     } else if (params.configId === "mode") {
       if (typeof params.value !== "string") throw RequestError.invalidParams("mode value must be a string")
       const availableModes = await this.loadAvailableModes(session.cwd)
@@ -1341,15 +1358,21 @@ export class Agent implements ACPAgent {
     const updatedSession = this.sessionManager.get(session.id)
     const model = updatedSession.model ?? (await defaultModel(this.config, session.cwd))
     const availableVariants = modelVariantsFromProviders(entries, model)
-    const currentModelId = formatModelIdWithVariant(model, updatedSession.variant, availableVariants, true)
-    const availableModels = buildAvailableModels(entries, { includeVariants: true })
+    const currentModelId = formatModelIdWithVariant(model, updatedSession.variant, availableVariants, false)
+    const availableModels = buildAvailableModels(entries)
     const modeState = await this.resolveModeState(session.cwd, session.id)
     const modes = modeState.currentModeId
       ? { availableModes: modeState.availableModes, currentModeId: modeState.currentModeId }
       : undefined
 
     return {
-      configOptions: buildConfigOptions({ currentModelId, availableModels, modes }),
+      configOptions: buildConfigOptions({
+        currentModelId,
+        availableModels,
+        currentVariant: updatedSession.variant,
+        availableVariants,
+        modes,
+      }),
     }
   }
 
@@ -1546,6 +1569,37 @@ export class Agent implements ACPAgent {
       { throwOnError: true },
     )
   }
+
+  private async loadSessionMessages(directory: string, sessionId: string, limit?: number) {
+    return this.sdk.session
+      .messages(
+        {
+          sessionID: sessionId,
+          directory,
+          limit,
+        },
+        { throwOnError: true },
+      )
+      .then((x) => x.data)
+      .catch((error) => {
+        log.error("unexpected error when fetching message", { error })
+        return undefined
+      })
+  }
+
+  private restoreSessionStateFromMessages(sessionId: string, messages: SessionMessageResponse[] | undefined) {
+    const lastUser = messages?.findLast((message) => message.info.role === "user")?.info
+    if (lastUser?.role !== "user") return
+
+    this.sessionManager.setModel(sessionId, {
+      providerID: ProviderID.make(lastUser.model.providerID),
+      modelID: ModelID.make(lastUser.model.modelID),
+    })
+    this.sessionManager.setVariant(sessionId, lastUser.model.variant)
+    if (lastUser.agent) {
+      this.sessionManager.setMode(sessionId, lastUser.agent)
+    }
+  }
 }
 
 function toToolKind(toolName: string): ToolKind {
@@ -1629,11 +1683,11 @@ async function defaultModel(config: ACPConfig, cwd?: string): Promise<{ provider
 
   if (specified && !providers.length) return specified
 
+  const lastUsed = await lastUsedModel(sdk, directory, providers)
+  if (lastUsed) return lastUsed
+
   const opencodeProvider = providers.find((p) => p.id === "opencode")
   if (opencodeProvider) {
-    if (opencodeProvider.models["big-pickle"]) {
-      return { providerID: ProviderID.opencode, modelID: ModelID.make("big-pickle") }
-    }
     const [best] = Provider.sort(Object.values(opencodeProvider.models))
     if (best) {
       return {
@@ -1653,8 +1707,38 @@ async function defaultModel(config: ACPConfig, cwd?: string): Promise<{ provider
   }
 
   if (specified) return specified
+  throw new Error("No models available")
+}
 
-  return { providerID: ProviderID.opencode, modelID: ModelID.make("big-pickle") }
+async function lastUsedModel(
+  sdk: OpencodeClient,
+  directory: string,
+  providers: Array<{ id: string; models: Record<string, unknown> }>,
+): Promise<{ providerID: ProviderID; modelID: ModelID } | undefined> {
+  const session = await sdk.session
+    .list({ directory, roots: true, limit: 1 }, { throwOnError: true })
+    .then((x) => x.data?.[0])
+    .catch((error) => {
+      log.error("failed to list sessions for default model", { error })
+      return undefined
+    })
+  if (!session) return
+
+  const lastUser = await sdk.session
+    .messages({ sessionID: session.id, directory, limit: 20 }, { throwOnError: true })
+    .then((x) => x.data?.findLast((message) => message.info.role === "user")?.info)
+    .catch((error) => {
+      log.error("failed to load session messages for default model", { error, sessionID: session.id })
+      return undefined
+    })
+  if (lastUser?.role !== "user") return
+
+  const provider = providers.find((entry) => entry.id === lastUser.model.providerID)
+  if (!provider?.models[lastUser.model.modelID]) return
+  return {
+    providerID: ProviderID.make(lastUser.model.providerID),
+    modelID: ModelID.make(lastUser.model.modelID),
+  }
 }
 
 function parseUri(
@@ -1757,8 +1841,14 @@ function formatModelIdWithVariant(
   includeVariant: boolean,
 ) {
   const base = `${model.providerID}/${model.modelID}`
-  if (!includeVariant || !variant || !availableVariants.includes(variant)) return base
-  return `${base}/${variant}`
+  if (!includeVariant || availableVariants.length === 0) return base
+  const selectedVariant =
+    variant && availableVariants.includes(variant)
+      ? variant
+      : availableVariants.includes(DEFAULT_VARIANT_VALUE)
+        ? DEFAULT_VARIANT_VALUE
+        : availableVariants[0]
+  return `${base}/${selectedVariant}`
 }
 
 function buildVariantMeta(input: {
@@ -1810,6 +1900,8 @@ function parseModelSelection(
 function buildConfigOptions(input: {
   currentModelId: string
   availableModels: ModelOption[]
+  currentVariant?: string
+  availableVariants?: string[]
   modes?: { availableModes: ModeOption[]; currentModeId: string } | undefined
 }): SessionConfigOption[] {
   const options: SessionConfigOption[] = [
@@ -1822,6 +1914,22 @@ function buildConfigOptions(input: {
       options: input.availableModels.map((m) => ({ value: m.modelId, name: m.name })),
     },
   ]
+  if (input.availableVariants?.length) {
+    options.push({
+      id: "effort",
+      name: "Effort",
+      description: "Available effort levels for this model",
+      category: "thought_level",
+      type: "select",
+      currentValue:
+        input.currentVariant && input.availableVariants.includes(input.currentVariant)
+          ? input.currentVariant
+          : input.availableVariants.includes(DEFAULT_VARIANT_VALUE)
+            ? DEFAULT_VARIANT_VALUE
+            : input.availableVariants[0],
+      options: input.availableVariants.map((variant) => ({ value: variant, name: formatVariantName(variant) })),
+    })
+  }
   if (input.modes) {
     options.push({
       id: "mode",
@@ -1839,4 +1947,11 @@ function buildConfigOptions(input: {
   return options
 }
 
+function formatVariantName(variant: string) {
+  return variant
+    .split(/[_-]/)
+    .map((part) => (part ? part.charAt(0).toUpperCase() + part.slice(1) : part))
+    .join(" ")
+}
+
 export * as ACP from "./agent"
diff --git a/packages/opencode/src/acp/session.ts b/packages/opencode/src/acp/session.ts
index d932b65701..cc1ed0be30 100644
--- a/packages/opencode/src/acp/session.ts
+++ b/packages/opencode/src/acp/session.ts
@@ -113,4 +113,10 @@ export class ACPSessionManager {
     this.sessions.set(sessionId, session)
     return session
   }
+
+  remove(sessionId: string): ACPSessionState | undefined {
+    const session = this.sessions.get(sessionId)
+    this.sessions.delete(sessionId)
+    return session
+  }
 }
diff --git a/packages/opencode/src/cli/cmd/acp.ts b/packages/opencode/src/cli/cmd/acp.ts
index e24262307c..b3b7df486b 100644
--- a/packages/opencode/src/cli/cmd/acp.ts
+++ b/packages/opencode/src/cli/cmd/acp.ts
@@ -22,7 +22,7 @@ export const AcpCommand = effectCmd({
   },
   handler: Effect.fn("Cli.acp")(function* (args) {
     process.env.OPENCODE_CLIENT = "acp"
-    const opts = yield* Effect.promise(() => resolveNetworkOptions(args))
+    const opts = yield* resolveNetworkOptions(args)
     const server = yield* Effect.promise(() => Server.listen(opts))
 
     const sdk = createOpencodeClient({
diff --git a/packages/opencode/src/cli/cmd/agent.ts b/packages/opencode/src/cli/cmd/agent.ts
index 313858e641..beba0c3773 100644
--- a/packages/opencode/src/cli/cmd/agent.ts
+++ b/packages/opencode/src/cli/cmd/agent.ts
@@ -84,7 +84,7 @@ const AgentCreateCommand = effectCmd({
       // Determine scope/path
       let targetPath: string
       if (cliPath) {
-        targetPath = path.join(cliPath, "agent")
+        targetPath = path.join(cliPath, "agents")
       } else {
         let scope: "global" | "project" = "global"
         if (project.vcs === "git") {
@@ -106,7 +106,7 @@ const AgentCreateCommand = effectCmd({
           if (prompts.isCancel(scopeResult)) throw new UI.CancelledError()
           scope = scopeResult
         }
-        targetPath = path.join(scope === "global" ? Global.Path.config : path.join(ctx.worktree, ".bcode"), "agent")
+        targetPath = path.join(scope === "global" ? Global.Path.config : path.join(ctx.worktree, ".bcode"), "agents")
       }
 
       // Get description
diff --git a/packages/opencode/src/cli/cmd/serve.ts b/packages/opencode/src/cli/cmd/serve.ts
index 2727794efc..5829f72a98 100644
--- a/packages/opencode/src/cli/cmd/serve.ts
+++ b/packages/opencode/src/cli/cmd/serve.ts
@@ -15,7 +15,7 @@ export const ServeCommand = effectCmd({
     if (!Flag.OPENCODE_SERVER_PASSWORD) {
       console.log("Warning: OPENCODE_SERVER_PASSWORD is not set; server is unsecured.")
     }
-    const opts = yield* Effect.promise(() => resolveNetworkOptions(args))
+    const opts = yield* resolveNetworkOptions(args)
     const server = yield* Effect.promise(() => Server.listen(opts))
     console.log(`opencode server listening on http://${server.hostname}:${server.port}`)
 
diff --git a/packages/opencode/src/cli/cmd/session.ts b/packages/opencode/src/cli/cmd/session.ts
index 08c0df929c..1240fa92ce 100644
--- a/packages/opencode/src/cli/cmd/session.ts
+++ b/packages/opencode/src/cli/cmd/session.ts
@@ -9,6 +9,7 @@ import { Locale } from "@/util/locale"
 import { Flag } from "@opencode-ai/core/flag/flag"
 import { Filesystem } from "@/util/filesystem"
 import { Process } from "@/util/process"
+import { NotFoundError } from "@/storage/storage"
 import { EOL } from "os"
 import path from "path"
 import { which } from "../../util/which"
@@ -59,9 +60,9 @@ export const SessionDeleteCommand = effectCmd({
   handler: Effect.fn("Cli.session.delete")(function* (args) {
     const svc = yield* Session.Service
     const sessionID = SessionID.make(args.sessionID)
-    // Match legacy try/catch — Session.get surfaces NotFoundError as a defect.
-    yield* svc.get(sessionID).pipe(Effect.catchCause(() => fail(`Session not found: ${args.sessionID}`)))
-    yield* svc.remove(sessionID)
+    yield* svc
+      .remove(sessionID)
+      .pipe(Effect.catchIf(NotFoundError.isInstance, () => fail(`Session not found: ${args.sessionID}`)))
     UI.println(UI.Style.TEXT_SUCCESS_BOLD + `Session ${args.sessionID} deleted` + UI.Style.TEXT_NORMAL)
   }),
 })
diff --git a/packages/opencode/src/cli/cmd/tui/component/dialog-provider.tsx b/packages/opencode/src/cli/cmd/tui/component/dialog-provider.tsx
index d6cbda4133..e12492a2d0 100644
--- a/packages/opencode/src/cli/cmd/tui/component/dialog-provider.tsx
+++ b/packages/opencode/src/cli/cmd/tui/component/dialog-provider.tsx
@@ -25,6 +25,60 @@ const PROVIDER_PRIORITY: Record<string, number> = {
   google: 5,
 }
 
+const CUSTOM_PROVIDER_OPTION_VALUE = "__opencode_custom_provider__"
+const CUSTOM_PROVIDER_ID = /^[a-z0-9][a-z0-9-_]*$/
+
+type ProviderOptionBase = {
+  title: string
+  value: string
+  description?: string
+  category: string
+}
+
+type ProviderOption =
+  | (ProviderOptionBase & {
+      type: "provider"
+      providerID: string
+    })
+  | (ProviderOptionBase & {
+      type: "custom"
+    })
+
+export function providerOptions(list: { id: string; name: string }[]): ProviderOption[] {
+  return [
+    ...pipe(
+      list,
+      sortBy((x) => PROVIDER_PRIORITY[x.id] ?? 99),
+      map((provider) => ({
+        type: "provider" as const,
+        title: provider.name,
+        value: provider.id,
+        providerID: provider.id,
+        description: {
+          opencode: "(Recommended)",
+          anthropic: "(API key)",
+          openai: "(ChatGPT Plus/Pro or API key)",
+          "opencode-go": "Low cost subscription for everyone",
+        }[provider.id],
+        category: provider.id in PROVIDER_PRIORITY ? "Popular" : "Providers",
+      })),
+    ),
+    {
+      type: "custom",
+      title: "Other",
+      value: CUSTOM_PROVIDER_OPTION_VALUE,
+      description: "Custom provider",
+      category: "Providers",
+    },
+  ]
+}
+
+export function normalizeCustomProviderID(value: string) {
+  const providerID = value.trim().replace(/^@ai-sdk\//, "")
+  if (!CUSTOM_PROVIDER_ID.test(providerID)) return
+  return providerID
+}
+
 export function createDialogProviderOptions() {
   const sync = useSync()
   const dialog = useDialog()
@@ -32,30 +86,62 @@ export function createDialogProviderOptions() {
   const toast = useToast()
   const { theme } = useTheme()
   const onboarded = useConnected()
+
+  async function promptCustomProviderID(): Promise<string | undefined> {
+    const value = await DialogPrompt.show(dialog, "Other", {
+      placeholder: "Provider id",
+      description: () => (
+        <text fg={theme.textMuted}>
+          This only stores a credential. Configure the provider in opencode.json to use it.
+        </text>
+      ),
+    })
+    if (value === null) return
+
+    const providerID = normalizeCustomProviderID(value)
+    if (providerID) return providerID
+
+    toast.show({
+      variant: "error",
+      message:
+        "Provider ids must start with a lowercase letter or number and only use lowercase letters, numbers, hyphens, and underscores",
+    })
+    return promptCustomProviderID()
+  }
+
   const options = createMemo(() => {
     return pipe(
-      sync.data.provider_next.all,
-      sortBy((x) => PROVIDER_PRIORITY[x.id] ?? 99),
+      providerOptions(sync.data.provider_next.all),
       map((provider) => {
-        const consoleManaged = isConsoleManagedProvider(sync.data.console_state.consoleManagedProviders, provider.id)
-        const connected = sync.data.provider_next.connected.includes(provider.id)
+        if (provider.type === "custom") {
+          return {
+            title: provider.title,
+            value: provider.value,
+            description: provider.description,
+            category: provider.category,
+            async onSelect() {
+              const providerID = await promptCustomProviderID()
+              if (!providerID) return
+              return dialog.replace(() => <ApiMethod providerID={providerID} title="API key" custom />)
+            },
+          }
+        }
+
+        const providerID = provider.providerID
+        const consoleManaged = isConsoleManagedProvider(sync.data.console_state.consoleManagedProviders, providerID)
+        const connected = sync.data.provider_next.connected.includes(providerID)
 
         return {
-          title: provider.name,
-          value: provider.id,
-          description: {
-            opencode: "(Recommended)",
-            anthropic: "(API key)",
-            openai: "(ChatGPT Plus/Pro or API key)",
-            "opencode-go": "Low cost subscription for everyone",
-          }[provider.id],
+          title: provider.title,
+          value: provider.value,
+          description: provider.description,
           footer: consoleManaged ? sync.data.console_state.activeOrgName : undefined,
-          category: provider.id in PROVIDER_PRIORITY ? "Popular" : "Other",
+          category: provider.category,
           gutter: connected && onboarded() ? () => <text fg={theme.success}>✓</text> : undefined,
           async onSelect() {
             if (consoleManaged) return
 
-            const methods = sync.data.provider_auth[provider.id] ?? [
+            const methods = sync.data.provider_auth[providerID] ?? [
               {
                 type: "api",
                 label: "API key",
@@ -93,7 +179,7 @@ export function createDialogProviderOptions() {
               }
 
               const result = await sdk.client.provider.oauth.authorize({
-                providerID: provider.id,
+                providerID,
                 method: index,
                 inputs,
               })
@@ -107,22 +193,12 @@ export function createDialogProviderOptions() {
               }
               if (result.data?.method === "code") {
                 dialog.replace(() => (
-                  <CodeMethod
-                    providerID={provider.id}
-                    title={method.label}
-                    index={index}
-                    authorization={result.data!}
-                  />
+                  <CodeMethod providerID={providerID} title={method.label} index={index} authorization={result.data!} />
                 ))
               }
               if (result.data?.method === "auto") {
                 dialog.replace(() => (
-                  <AutoMethod
-                    providerID={provider.id}
-                    title={method.label}
-                    index={index}
-                    authorization={result.data!}
-                  />
+                  <AutoMethod providerID={providerID} title={method.label} index={index} authorization={result.data!} />
                 ))
               }
             }
@@ -134,7 +210,7 @@ export function createDialogProviderOptions() {
                 metadata = value
               }
               return dialog.replace(() => (
-                <ApiMethod providerID={provider.id} title={method.label} metadata={metadata} />
+                <ApiMethod providerID={providerID} title={method.label} metadata={metadata} />
               ))
             }
           },
@@ -256,11 +332,13 @@ interface ApiMethodProps {
   providerID: string
   title: string
   metadata?: Record<string, string>
+  custom?: boolean
 }
 function ApiMethod(props: ApiMethodProps) {
   const dialog = useDialog()
   const sdk = useSDK()
   const sync = useSync()
+  const toast = useToast()
   const { theme } = useTheme()
 
   return (
@@ -305,6 +383,14 @@ function ApiMethod(props: ApiMethodProps) {
         })
         await sdk.client.instance.dispose()
         await sync.bootstrap()
+        if (props.custom && !sync.data.provider_next.all.some((provider) => provider.id === props.providerID)) {
+          toast.show({
+            variant: "info",
+            message: `Saved credential for ${props.providerID}. Configure it in opencode.json to use it.`,
+          })
+          dialog.clear()
+          return
+        }
         dialog.replace(() => <DialogModel providerID={props.providerID} />)
       }}
     />
diff --git a/packages/opencode/src/cli/cmd/tui/component/dialog-session-list.tsx b/packages/opencode/src/cli/cmd/tui/component/dialog-session-list.tsx
index 09d952ef81..e8dbaee394 100644
--- a/packages/opencode/src/cli/cmd/tui/component/dialog-session-list.tsx
+++ b/packages/opencode/src/cli/cmd/tui/component/dialog-session-list.tsx
@@ -70,8 +70,10 @@ export function DialogSessionList() {
         sync,
         project,
         toast,
+        sourceWorkspaceID: session.workspaceID,
         workspaceID,
         sessionID: session.id,
+        copyChanges: false,
         done: list,
       })
     }
diff --git a/packages/opencode/src/cli/cmd/tui/component/dialog-workspace-create.tsx b/packages/opencode/src/cli/cmd/tui/component/dialog-workspace-create.tsx
index ad40637575..d7e212ab15 100644
--- a/packages/opencode/src/cli/cmd/tui/component/dialog-workspace-create.tsx
+++ b/packages/opencode/src/cli/cmd/tui/component/dialog-workspace-create.tsx
@@ -3,10 +3,13 @@ import { useDialog } from "@tui/ui/dialog"
 import { DialogSelect, type DialogSelectOption } from "@tui/ui/dialog-select"
 import { useSync } from "@tui/context/sync"
 import { useProject } from "@tui/context/project"
+import { useRoute } from "@tui/context/route"
 import { createMemo, createSignal, onMount } from "solid-js"
 import { errorMessage } from "@/util/error"
 import { useSDK } from "../context/sdk"
 import { useToast } from "../ui/toast"
+import { DialogAlert } from "../ui/dialog-alert"
+import { DialogWorkspaceFileChanges } from "./dialog-workspace-file-changes"
 
 type Adapter = {
   type: string
@@ -33,6 +36,30 @@ export type WorkspaceSelection =
 type WorkspaceSelectValue = WorkspaceSelection | { type: "existing-list" }
 type ExistingWorkspaceSelectValue = { workspace: Workspace }
 
+export function recentConnectedWorkspaces<WorkspaceInfo extends { id: string }>(input: {
+  sessions: readonly { workspaceID?: string; time: { updated: number } }[]
+  get: (workspaceID: string) => WorkspaceInfo | undefined
+  status: (workspaceID: string) => string | undefined
+  limit?: number
+  omitWorkspaceID?: string
+}) {
+  const workspaces = input.sessions
+    .toSorted((a, b) => b.time.updated - a.time.updated)
+    .flatMap((session) => {
+      const workspace = session.workspaceID ? input.get(session.workspaceID) : undefined
+      return workspace && input.status(workspace.id) === "connected" ? [workspace] : []
+    })
+    .filter((workspace) => workspace.id !== input.omitWorkspaceID)
+    .filter((workspace, index, list) => list.findIndex((item) => item.id === workspace.id) === index)
+  const recent = workspaces.slice(0, input.limit ?? 3)
+
+  return { recent, hasMore: recent.length < workspaces.length }
+}
+
+export function warpReminderText(dir: string) {
+  return `<system-reminder>The user has changed the current working directory to "${dir}". This is still the same project but at a possibly new location; take this into account when working with any files from now on.</system-reminder>`
+}
+
 async function loadWorkspaceAdapters(input: {
   sdk: ReturnType<typeof useSDK>
   sync: ReturnType<typeof useSync>
@@ -71,17 +98,29 @@ export async function warpWorkspaceSession(input: {
   sync: ReturnType<typeof useSync>
   project: ReturnType<typeof useProject>
   toast: ReturnType<typeof useToast>
+  sourceWorkspaceID?: string
   workspaceID: string | null
   sessionID: string
+  copyChanges: boolean
   done?: () => void
 }): Promise<boolean> {
   const result = await input.sdk.client.experimental.workspace
     .warp({
-      id: input.workspaceID ?? undefined,
+      id: input.workspaceID,
       sessionID: input.sessionID,
+      copyChanges: input.copyChanges,
     })
     .catch(() => undefined)
   if (!result?.data) {
+    if (result?.error?.name === "VcsApplyError") {
+      await DialogAlert.show(
+        input.dialog,
+        "Unable to Warp Session",
+        "Unable to apply file changes to this workspace. It has existing changes that conflict or is based off a different branch. Session has not been warped.",
+      )
+      return false
+    }
+
     input.toast.show({
       message: `Failed to warp session: ${errorMessage(result?.error ?? "no response")}`,
       variant: "error",
@@ -93,24 +132,59 @@ export async function warpWorkspaceSession(input: {
 
   await input.sync.bootstrap({ fatal: false }).catch(() => undefined)
 
+  const dir = input.project.instance.directory() || input.sync.path.directory
+  if (dir) {
+    await input.sdk.client.session
+      .promptAsync({
+        sessionID: input.sessionID,
+        workspace: input.workspaceID ?? undefined,
+        noReply: true,
+        parts: [
+          {
+            type: "text",
+            text: warpReminderText(dir),
+            synthetic: true,
+          },
+        ],
+      })
+      .catch(() => undefined)
+  }
+
   await Promise.all([input.project.workspace.sync(), input.sync.session.refresh()])
 
-  input.done?.()
-  if (input.done) return true
+  if (input.done) {
+    input.done()
+    return true
+  }
   input.dialog.clear()
   return true
 }
 
+export async function confirmWorkspaceFileChanges(input: {
+  dialog: ReturnType<typeof useDialog>
+  sdk: ReturnType<typeof useSDK>
+  sourceWorkspaceID?: string
+}) {
+  const status = await input.sdk.client.vcs.status({ workspace: input.sourceWorkspaceID }).catch(() => undefined)
+  const fileChangeChoice = status?.data?.length
+    ? await DialogWorkspaceFileChanges.show(input.dialog, status.data)
+    : "no"
+  if (!fileChangeChoice) return
+  return fileChangeChoice === "yes"
+}
+
 export function DialogWorkspaceSelect(props: {
   adapters?: Adapter[]
   onSelect: (selection: WorkspaceSelection) => Promise<void> | void
 }) {
   const dialog = useDialog()
   const project = useProject()
+  const route = useRoute()
   const sync = useSync()
   const sdk = useSDK()
   const toast = useToast()
   const [adapters, setAdapters] = createSignal<Adapter[] | undefined>(props.adapters)
+  const omittedWorkspaceID = createMemo(() => (route.data.type === "session" ? project.workspace.current() : undefined))
 
   onMount(() => {
     dialog.setSize("medium")
@@ -125,15 +199,12 @@ export function DialogWorkspaceSelect(props: {
   const options = createMemo<DialogSelectOption<WorkspaceSelectValue>[]>(() => {
     const list = adapters()
     if (!list) return []
-    const recent = sync.data.session
-      .toSorted((a, b) => b.time.updated - a.time.updated)
-      .flatMap((session) => (session.workspaceID ? [session.workspaceID] : []))
-      .filter((workspaceID, index, list) => list.indexOf(workspaceID) === index)
-      .slice(0, 3)
-      .flatMap((workspaceID) => {
-        const workspace = project.workspace.get(workspaceID)
-        return workspace ? [workspace] : []
-      })
+    const { recent, hasMore } = recentConnectedWorkspaces({
+      sessions: sync.data.session,
+      get: project.workspace.get,
+      status: project.workspace.status,
+      omitWorkspaceID: omittedWorkspaceID(),
+    })
     return [
       ...list.map((adapter) => ({
         title: adapter.name,
@@ -158,12 +229,16 @@ export function DialogWorkspaceSelect(props: {
         },
         category: "Choose workspace",
       })),
-      {
-        title: "View all workspaces",
-        value: { type: "existing-list" as const },
-        description: "Choose from all workspaces",
-        category: "Choose workspace",
-      },
+      ...(hasMore
+        ? [
+            {
+              title: "View all workspaces",
+              value: { type: "existing-list" as const },
+              description: "Choose from all workspaces",
+              category: "Choose workspace",
+            },
+          ]
+        : []),
     ]
   })
 
@@ -189,19 +264,25 @@ export function DialogWorkspaceSelect(props: {
           return
         }
 
-        dialog.replace(() => <DialogExistingWorkspaceSelect onSelect={props.onSelect} />)
+        dialog.replace(() => (
+          <DialogExistingWorkspaceSelect omitWorkspaceID={omittedWorkspaceID()} onSelect={props.onSelect} />
+        ))
       }}
     />
   )
 }
 
-function DialogExistingWorkspaceSelect(props: { onSelect: (selection: WorkspaceSelection) => Promise<void> | void }) {
+function DialogExistingWorkspaceSelect(props: {
+  omitWorkspaceID?: string
+  onSelect: (selection: WorkspaceSelection) => Promise<void> | void
+}) {
   const project = useProject()
 
   const options = createMemo<DialogSelectOption<ExistingWorkspaceSelectValue>[]>(() =>
     project.workspace
       .list()
       .filter((workspace) => project.workspace.status(workspace.id) === "connected")
+      .filter((workspace) => workspace.id !== props.omitWorkspaceID)
       .map((workspace: Workspace) => ({
         title: workspace.name,
         description: `(${workspace.type})`,
diff --git a/packages/opencode/src/cli/cmd/tui/component/dialog-workspace-file-changes.tsx b/packages/opencode/src/cli/cmd/tui/component/dialog-workspace-file-changes.tsx
new file mode 100644
index 0000000000..b2cb20630c
--- /dev/null
+++ b/packages/opencode/src/cli/cmd/tui/component/dialog-workspace-file-changes.tsx
@@ -0,0 +1,138 @@
+import { TextAttributes } from "@opentui/core"
+import { useKeyboard } from "@opentui/solid"
+import type { VcsFileStatus } from "@opencode-ai/sdk/v2"
+import { createMemo, For } from "solid-js"
+import { createStore } from "solid-js/store"
+import { Locale } from "@/util/locale"
+import { useTheme } from "../context/theme"
+import { useTuiConfig } from "../context/tui-config"
+import { useDialog, type DialogContext } from "../ui/dialog"
+import { getScrollAcceleration } from "../util/scroll"
+
+const options = ["no", "yes"] as const
+
+export type WorkspaceFileChangesChoice = (typeof options)[number]
+
+function statusLabel(status: VcsFileStatus["status"]) {
+  if (status === "added") return "A"
+  if (status === "deleted") return "D"
+  return "M"
+}
+
+function changeCountWidth(file: VcsFileStatus) {
+  // The "plus 2" is for spaces
+  return `${file.additions ? `+${file.additions}` : ""}${file.deletions ? ` -${file.deletions}` : ""}`.length + 2
+}
+
+export function DialogWorkspaceFileChanges(props: {
+  files: VcsFileStatus[]
+  onSelect: (choice: WorkspaceFileChangesChoice) => void
+}) {
+  const dialog = useDialog()
+  const { theme } = useTheme()
+  const tuiConfig = useTuiConfig()
+  const scrollAcceleration = createMemo(() => getScrollAcceleration(tuiConfig))
+  const [store, setStore] = createStore({ active: "yes" as WorkspaceFileChangesChoice })
+  const height = createMemo(() => Math.min(props.files.length, 8))
+  const fileNameWidth = createMemo(() => 48 - Math.max(Math.max(7, ...props.files.map(changeCountWidth)) - 7, 0))
+
+  function confirm() {
+    props.onSelect(store.active)
+    dialog.clear()
+  }
+
+  useKeyboard((evt) => {
+    if (evt.name === "return") {
+      evt.preventDefault()
+      evt.stopPropagation()
+      confirm()
+      return
+    }
+    if (evt.name === "left") {
+      evt.preventDefault()
+      evt.stopPropagation()
+      const index = options.indexOf(store.active)
+      setStore("active", options[Math.max(index - 1, 0)])
+      return
+    }
+    if (evt.name === "right") {
+      evt.preventDefault()
+      evt.stopPropagation()
+      const index = options.indexOf(store.active)
+      setStore("active", options[Math.min(index + 1, options.length - 1)])
+    }
+  })
+
+  return (
+    <box gap={1}>
+      <box flexDirection="row" justifyContent="space-between" paddingLeft={2} paddingRight={2}>
+        <text attributes={TextAttributes.BOLD} fg={theme.text}>
+          File Changes Found
+        </text>
+        <text fg={theme.textMuted} onMouseUp={() => dialog.clear()}>
+          esc
+        </text>
+      </box>
+      <scrollbox
+        height={height()}
+        backgroundColor={theme.backgroundElement}
+        scrollbarOptions={{ visible: false }}
+        scrollAcceleration={scrollAcceleration()}
+      >
+        <For each={props.files}>
+          {(item) => (
+            <box flexDirection="row" justifyContent="space-between" paddingLeft={2} paddingRight={2}>
+              <box flexDirection="row" minWidth={0} flexShrink={1}>
+                <box width={2} flexShrink={0}>
+                  <text fg={theme.textMuted}>{statusLabel(item.status)}</text>
+                </box>
+                <text fg={theme.textMuted} wrapMode="none">
+                  {Locale.truncateLeft(item.file, fileNameWidth())}
+                </text>
+              </box>
+              <box flexDirection="row" gap={1} minWidth={7} flexShrink={0} justifyContent="flex-end">
+                <text>
+                  {" "}
+                  {item.additions ? <span style={{ fg: theme.diffAdded }}>+{item.additions}</span> : null}
+                  {item.deletions ? <span style={{ fg: theme.diffRemoved }}> -{item.deletions}</span> : null}
+                </text>
+              </box>
+            </box>
+          )}
+        </For>
+      </scrollbox>
+      <box paddingLeft={2} paddingRight={2}>
+        <text fg={theme.textMuted} wrapMode="word">
+          Do you want to apply these changes after warping?
+        </text>
+      </box>
+      <box flexDirection="row" justifyContent="flex-end" paddingLeft={2} paddingRight={2} paddingBottom={1}>
+        <For each={options}>
+          {(item) => (
+            <box
+              paddingLeft={2}
+              paddingRight={2}
+              backgroundColor={item === store.active ? theme.primary : undefined}
+              onMouseUp={() => {
+                setStore("active", item)
+                props.onSelect(item)
+                dialog.clear()
+              }}
+            >
+              <text fg={item === store.active ? theme.selectedListItemText : theme.textMuted}>{item}</text>
+            </box>
+          )}
+        </For>
+      </box>
+    </box>
+  )
+}
+
+DialogWorkspaceFileChanges.show = (dialog: DialogContext, files: VcsFileStatus[]) => {
+  return new Promise<WorkspaceFileChangesChoice | undefined>((resolve) => {
+    dialog.replace(
+      () => <DialogWorkspaceFileChanges files={files} onSelect={resolve} />,
+      () => resolve(undefined),
+    )
+  })
+}
diff --git a/packages/opencode/src/cli/cmd/tui/component/prompt/index.tsx b/packages/opencode/src/cli/cmd/tui/component/prompt/index.tsx
index 74332c77be..73ef5477e9 100644
--- a/packages/opencode/src/cli/cmd/tui/component/prompt/index.tsx
+++ b/packages/opencode/src/cli/cmd/tui/component/prompt/index.tsx
@@ -42,7 +42,12 @@ import { useKV } from "../../context/kv"
 import { createFadeIn } from "../../util/signal"
 import { useTextareaKeybindings } from "../textarea-keybindings"
 import { DialogSkill } from "../dialog-skill"
-import { openWorkspaceSelect, warpWorkspaceSession, type WorkspaceSelection } from "../dialog-workspace-create"
+import {
+  confirmWorkspaceFileChanges,
+  openWorkspaceSelect,
+  warpWorkspaceSession,
+  type WorkspaceSelection,
+} from "../dialog-workspace-create"
 import { DialogWorkspaceUnavailable } from "../dialog-workspace-unavailable"
 import { useArgs } from "@tui/context/args"
 import { Flag } from "@opencode-ai/core/flag/flag"
@@ -173,8 +178,7 @@ export function Prompt(props: PromptProps) {
     if (!file) return
     return Locale.truncateMiddle(file, Math.max(12, Math.min(48, Math.floor(dimensions().width / 3))))
   })
-  const [editorContextHover, setEditorContextHover] = createSignal(false)
-  let lastSubmittedEditorSelectionKey: string | undefined
+  const editorContextLabelState = createMemo(() => editor.labelState())
   const [auto, setAuto] = createSignal<AutocompleteRef>()
   const [workspaceSelection, setWorkspaceSelection] = createSignal<WorkspaceSelection>()
   const [workspaceCreating, setWorkspaceCreating] = createSignal(false)
@@ -182,6 +186,7 @@ export function Prompt(props: PromptProps) {
   const [warpNotice, setWarpNotice] = createSignal<string>()
   const currentProviderLabel = createMemo(() => local.model.parsed().provider)
   const hasRightContent = createMemo(() => Boolean(props.right))
+  const defaultWorkspaceID = createMemo(() => props.workspaceID ?? project.workspace.current())
 
   function selectWorkspace(selection: WorkspaceSelection | undefined) {
     setWorkspaceSelection(selection)
@@ -230,6 +235,9 @@ export function Prompt(props: PromptProps) {
       if (selection.type === "new") void createWorkspace(selection)
       return
     }
+    const sourceWorkspaceID = project.workspace.current()
+    const copyChanges = await confirmWorkspaceFileChanges({ dialog, sdk, sourceWorkspaceID })
+    if (copyChanges === undefined) return
     selectWorkspace(selection)
     dialog.clear()
 
@@ -247,8 +255,10 @@ export function Prompt(props: PromptProps) {
       sync,
       project,
       toast,
+      sourceWorkspaceID,
       workspaceID: workspace.id,
       sessionID: props.sessionID,
+      copyChanges,
     })
     if (warped) showWarpNotice(workspace.name)
   }
@@ -861,14 +871,14 @@ export function Prompt(props: PromptProps) {
     if (sessionID == null) {
       const workspace = workspaceSelection()
       const workspaceID = iife(() => {
-        if (!workspace) return undefined
+        if (!workspace) return defaultWorkspaceID()
         if (workspace.type === "none") return undefined
         if (workspace.type === "existing") return workspace.workspaceID
         return undefined
       })
 
       const res = await sdk.client.session.create({
-        workspace: props.workspaceID,
+        workspace: workspaceID,
         agent: agent.name,
         model: {
           providerID: selectedModel.providerID,
@@ -916,9 +926,8 @@ export function Prompt(props: PromptProps) {
     // Capture mode before it gets reset
     const currentMode = store.mode
     const editorSelection = editorContext()
-    const currentEditorSelectionKey = editorSelectionKey(editorSelection)
     const editorParts =
-      editorSelection && currentEditorSelectionKey !== lastSubmittedEditorSelectionKey
+      editorSelection && editor.labelState() === "pending"
         ? [
             {
               id: PartID.ascending(),
@@ -996,7 +1005,7 @@ export function Prompt(props: PromptProps) {
           ],
         })
         .catch(() => {})
-      lastSubmittedEditorSelectionKey = currentEditorSelectionKey
+      if (editorParts.length > 0) editor.markSelectionSent()
     }
     history.append({
       ...store.prompt,
@@ -1011,13 +1020,15 @@ export function Prompt(props: PromptProps) {
     props.onSubmit?.()
 
     // temporary hack to make sure the message is sent
-    if (!props.sessionID)
+    if (!props.sessionID) {
+      if (editorParts.length > 0) editor.preserveSelectionFromNewSession()
       setTimeout(() => {
         route.navigate({
           type: "session",
           sessionID,
         })
       }, 50)
+    }
     input.clear()
     return true
   }
@@ -1145,7 +1156,17 @@ export function Prompt(props: PromptProps) {
     | undefined
   >(() => {
     const selected = workspaceSelection()
-    if (!selected) return
+    if (!selected) {
+      const workspaceID = defaultWorkspaceID()
+      if (props.sessionID || !workspaceID) return
+      const workspace = project.workspace.get(workspaceID)
+      return {
+        type: "existing",
+        workspaceType: workspace?.type ?? "unknown",
+        workspaceName: workspace?.name ?? workspaceID,
+        status: project.workspace.status(workspaceID) ?? "error",
+      }
+    }
     if (selected.type === "none") return
     if (props.sessionID && !workspaceCreating()) return
     if (selected.type === "new") {
@@ -1608,16 +1629,9 @@ export function Prompt(props: PromptProps) {
           </Switch>
           <Show when={status().type !== "retry"}>
             <box gap={2} flexDirection="row">
-              <Show when={editorFileLabelDisplay()}>
+              <Show when={editorContextLabelState() !== "none" ? editorFileLabelDisplay() : undefined}>
                 {(file) => (
-                  <text
-                    fg={theme.secondary}
-                    onMouseOver={() => setEditorContextHover(true)}
-                    onMouseOut={() => setEditorContextHover(false)}
-                    onMouseUp={dismissEditorContext}
-                  >
-                    {editorContextHover() ? `x ${file()}` : file()}
-                  </text>
+                  <text fg={editorContextLabelState() === "pending" ? theme.secondary : theme.textMuted}>{file()}</text>
                 )}
               </Show>
               <Switch>
diff --git a/packages/opencode/src/cli/cmd/tui/context/editor.ts b/packages/opencode/src/cli/cmd/tui/context/editor.ts
index 06dd6fd042..6d9e04cf84 100644
--- a/packages/opencode/src/cli/cmd/tui/context/editor.ts
+++ b/packages/opencode/src/cli/cmd/tui/context/editor.ts
@@ -87,6 +87,7 @@ const EditorServerInfoSchema = z.object({
 type JsonRpcMessage = z.infer<typeof JsonRpcMessageSchema>
 export type EditorSelection = z.infer<typeof EditorSelectionSchema>
 export type EditorMention = z.infer<typeof EditorMentionSchema>
+export type EditorLabelState = "pending" | "sent" | "none"
 type EditorServerInfo = z.infer<typeof EditorServerInfoSchema>
 
 type EditorConnection = {
@@ -111,10 +112,12 @@ export const { use: useEditorContext, provider: EditorContextProvider } = create
     const [store, setStore] = createStore<{
       status: "disabled" | "connecting" | "connected"
       selection: EditorSelection | undefined
+      selectionSent: boolean
       server: EditorServerInfo | undefined
     }>({
       status: "disabled",
       selection: undefined,
+      selectionSent: false,
       server: undefined,
     })
 
@@ -126,8 +129,24 @@ export const { use: useEditorContext, provider: EditorContextProvider } = create
     let zedSelection: Promise<void> | undefined
     let lastZedSelectionKey: string | undefined
     let directory = process.cwd()
+    let preserveSelectionOnReconnect = false
     const pending = new Map<number, string>()
 
+    const setSelection = (selection: EditorSelection | undefined) => {
+      const changed = editorSelectionKey(selection) !== editorSelectionKey(store.selection)
+      setStore("selection", selection)
+      if (changed) setStore("selectionSent", false)
+    }
+
+    const clearSelectionForReconnect = (options?: { resetZedSelectionKey?: boolean }) => {
+      if (preserveSelectionOnReconnect) {
+        preserveSelectionOnReconnect = false
+        return
+      }
+      if (options?.resetZedSelectionKey) lastZedSelectionKey = undefined
+      setSelection(undefined)
+    }
+
     const send = (payload: JsonRpcMessage) => {
       if (!socket || socket.readyState !== 1) return
       socket.send(JSON.stringify({ jsonrpc: "2.0", ...payload }))
@@ -158,7 +177,7 @@ export const { use: useEditorContext, provider: EditorContextProvider } = create
             const key = editorSelectionKey(selection)
             if (key !== lastZedSelectionKey) {
               lastZedSelectionKey = key
-              setStore("selection", selection)
+              setSelection(selection)
               setStore("status", selection ? "connected" : "disabled")
             }
           })
@@ -198,7 +217,7 @@ export const { use: useEditorContext, provider: EditorContextProvider } = create
         const selection =
           message.method === "selection_changed" ? EditorSelectionSchema.safeParse(message.params) : undefined
         if (selection?.success) {
-          setStore("selection", { ...selection.data, source: "websocket" })
+          setSelection({ ...selection.data, source: "websocket" })
           return
         }
 
@@ -252,12 +271,13 @@ export const { use: useEditorContext, provider: EditorContextProvider } = create
 
     const reconnectWithDirectory = (nextDirectory?: string) => {
       const resolved = nextDirectory || process.cwd()
-      if (directory === resolved) return
+      const sameDirectory = directory === resolved
+      clearSelectionForReconnect({ resetZedSelectionKey: !sameDirectory })
+      if (sameDirectory) return
 
       directory = resolved
       attempt = 0
       pending.clear()
-      lastZedSelectionKey = undefined
       if (reconnect) clearTimeout(reconnect)
       reconnect = undefined
       if (socket) {
@@ -266,7 +286,6 @@ export const { use: useEditorContext, provider: EditorContextProvider } = create
         current.close()
       }
       setStore("status", "disabled")
-      setStore("selection", undefined)
       setStore("server", undefined)
       connect()
     }
@@ -293,7 +312,19 @@ export const { use: useEditorContext, provider: EditorContextProvider } = create
       },
       clearSelection() {
         lastZedSelectionKey = undefined
-        setStore("selection", undefined)
+        zedSelection = undefined
+        setSelection(undefined)
+      },
+      preserveSelectionFromNewSession() {
+        preserveSelectionOnReconnect = true
+      },
+      markSelectionSent() {
+        if (!store.selection) return
+        setStore("selectionSent", true)
+      },
+      labelState(): EditorLabelState {
+        if (!store.selection) return "none"
+        return store.selectionSent ? "sent" : "pending"
       },
       onMention(listener: (mention: EditorMention) => void) {
         mentionListeners.add(listener)
@@ -303,7 +334,6 @@ export const { use: useEditorContext, provider: EditorContextProvider } = create
         return store.server
       },
       reconnect(directory?: string) {
-        setStore("selection", undefined)
         reconnectWithDirectory(directory)
       },
     }
diff --git a/packages/opencode/src/cli/cmd/tui/context/local.tsx b/packages/opencode/src/cli/cmd/tui/context/local.tsx
index 0b8c902c49..2958b573dd 100644
--- a/packages/opencode/src/cli/cmd/tui/context/local.tsx
+++ b/packages/opencode/src/cli/cmd/tui/context/local.tsx
@@ -397,23 +397,15 @@ export const { use: useLocal, provider: LocalProvider } = createSimpleContext({
       },
     }
 
-    // Automatically update model when agent changes
     createEffect(() => {
       const value = agent.current()
-      if (!value) return
-      if (value.model) {
-        if (isModelValid(value.model))
-          model.set({
-            providerID: value.model.providerID,
-            modelID: value.model.modelID,
-          })
-        else
-          toast.show({
-            variant: "warning",
-            message: `Agent ${value.name}'s configured model ${value.model.providerID}/${value.model.modelID} is not valid`,
-            duration: 3000,
-          })
-      }
+      if (!value?.model) return
+      if (isModelValid(value.model)) return
+      toast.show({
+        variant: "warning",
+        message: `Agent ${value.name}'s configured model ${value.model.providerID}/${value.model.modelID} is not valid`,
+        duration: 3000,
+      })
     })
 
     const result = {
diff --git a/packages/opencode/src/cli/cmd/tui/feature-plugins/home/tips-view.tsx b/packages/opencode/src/cli/cmd/tui/feature-plugins/home/tips-view.tsx
index 812a618813..488973cc71 100644
--- a/packages/opencode/src/cli/cmd/tui/feature-plugins/home/tips-view.tsx
+++ b/packages/opencode/src/cli/cmd/tui/feature-plugins/home/tips-view.tsx
@@ -94,7 +94,7 @@ const TIPS = [
   "Add {highlight}.md{/highlight} files to {highlight}.bcode/command/{/highlight} to define reusable custom prompts",
   "Use {highlight}$ARGUMENTS{/highlight}, {highlight}$1{/highlight}, {highlight}$2{/highlight} in custom commands for dynamic input",
   "Use backticks in commands to inject shell output (e.g., {highlight}`git status`{/highlight})",
-  "Add {highlight}.md{/highlight} files to {highlight}.bcode/agent/{/highlight} for specialized AI personas",
+  "Add {highlight}.md{/highlight} files to {highlight}.bcode/agents/{/highlight} for specialized AI personas",
   "Configure per-agent permissions for {highlight}edit{/highlight}, {highlight}bash{/highlight}, and {highlight}webfetch{/highlight} tools",
   'Use patterns like {highlight}"git *": "allow"{/highlight} for granular bash permissions',
   'Set {highlight}"rm -rf *": "deny"{/highlight} to block destructive commands',
diff --git a/packages/opencode/src/cli/cmd/tui/routes/home.tsx b/packages/opencode/src/cli/cmd/tui/routes/home.tsx
index 4c1cd1babd..43a52082be 100644
--- a/packages/opencode/src/cli/cmd/tui/routes/home.tsx
+++ b/packages/opencode/src/cli/cmd/tui/routes/home.tsx
@@ -1,5 +1,5 @@
 import { Prompt, type PromptRef } from "@tui/component/prompt"
-import { createEffect, createSignal } from "solid-js"
+import { createEffect, createSignal, onMount } from "solid-js"
 import { Logo } from "../component/logo"
 import { useProject } from "../context/project"
 import { useSync } from "../context/sync"
@@ -9,6 +9,7 @@ import { useRouteData } from "@tui/context/route"
 import { usePromptRef } from "../context/prompt"
 import { useLocal } from "../context/local"
 import { TuiPluginRuntime } from "@/cli/cmd/tui/plugin/runtime"
+import { useEditorContext } from "@tui/context/editor"
 
 let once = false
 const placeholder = {
@@ -24,8 +25,13 @@ export function Home() {
   const [ref, setRef] = createSignal<PromptRef | undefined>()
   const args = useArgs()
   const local = useLocal()
+  const editor = useEditorContext()
   let sent = false
 
+  onMount(() => {
+    editor.clearSelection()
+  })
+
   const bind = (r: PromptRef | undefined) => {
     setRef(r)
     promptRef.set(r)
diff --git a/packages/opencode/src/cli/cmd/web.ts b/packages/opencode/src/cli/cmd/web.ts
index b8e752356a..0e390ebb86 100644
--- a/packages/opencode/src/cli/cmd/web.ts
+++ b/packages/opencode/src/cli/cmd/web.ts
@@ -40,7 +40,7 @@ export const WebCommand = effectCmd({
     if (!Flag.OPENCODE_SERVER_PASSWORD) {
       UI.println(UI.Style.TEXT_WARNING_BOLD + "!  OPENCODE_SERVER_PASSWORD is not set; server is unsecured.")
     }
-    const opts = yield* Effect.promise(() => resolveNetworkOptions(args))
+    const opts = yield* resolveNetworkOptions(args)
     const server = yield* Effect.promise(() => Server.listen(opts))
     UI.empty()
     UI.println(UI.logo("  "))
@@ -72,7 +72,7 @@ export const WebCommand = effectCmd({
       }
 
       // Open localhost in browser
-      open(localhostUrl.toString()).catch(() => {})
+      open(localhostUrl).catch(() => {})
     } else {
       const displayUrl = server.url.toString()
       UI.println(UI.Style.TEXT_INFO_BOLD + "  Web interface:    ", UI.Style.TEXT_NORMAL, displayUrl)
diff --git a/packages/opencode/src/cli/network.ts b/packages/opencode/src/cli/network.ts
index b3c4611781..af0f589297 100644
--- a/packages/opencode/src/cli/network.ts
+++ b/packages/opencode/src/cli/network.ts
@@ -1,6 +1,6 @@
 import type { Argv, InferredOptionTypes } from "yargs"
 import { Config } from "@/config/config"
-import { AppRuntime } from "@/effect/app-runtime"
+import { Effect } from "effect"
 
 const options = {
   port: {
@@ -36,10 +36,10 @@ export type NetworkOptions = InferredOptionTypes<typeof options>
 export function withNetworkOptions<T>(yargs: Argv<T>) {
   return yargs.options(options)
 }
-export async function resolveNetworkOptions(args: NetworkOptions) {
-  const config = await AppRuntime.runPromise(Config.Service.use((cfg) => cfg.getGlobal()))
+export const resolveNetworkOptions = Effect.fn("Cli.resolveNetworkOptions")(function* (args: NetworkOptions) {
+  const config = yield* Config.Service.use((cfg) => cfg.getGlobal())
   return resolveNetworkOptionsNoConfig(args, config)
-}
+})
 
 export function resolveNetworkOptionsNoConfig(args: NetworkOptions, config?: Config.Info) {
   const portExplicitlySet = process.argv.includes("--port")
diff --git a/packages/opencode/src/config/config.ts b/packages/opencode/src/config/config.ts
index 72292aae1c..829f248e79 100644
--- a/packages/opencode/src/config/config.ts
+++ b/packages/opencode/src/config/config.ts
@@ -70,6 +70,36 @@ function normalizeLoadedConfig(data: unknown, source: string) {
   return copy
 }
 
+async function substituteWellKnownRemoteConfig(input: { value: unknown; dir: string; source: string }) {
+  if (!isRecord(input.value) || typeof input.value.url !== "string") return
+
+  const url = await ConfigVariable.substitute({
+    text: input.value.url,
+    type: "virtual",
+    dir: input.dir,
+    source: input.source,
+  })
+  const headers = isRecord(input.value.headers)
+    ? Object.fromEntries(
+        await Promise.all(
+          Object.entries(input.value.headers)
+            .filter((entry): entry is [string, string] => typeof entry[1] === "string")
+            .map(async ([key, value]) => [
+              key,
+              await ConfigVariable.substitute({
+                text: value,
+                type: "virtual",
+                dir: input.dir,
+                source: input.source,
+              }),
+            ]),
+        ),
+      )
+    : undefined
+
+  return { url, headers }
+}
+
 async function resolveLoadedPlugins<T extends { plugin?: ConfigPlugin.Spec[] }>(config: T, filepath: string) {
   if (!config.plugin) return config
   for (let i = 0; i < config.plugin.length; i++) {
@@ -501,8 +531,28 @@ export const layer = Layer.effect(
             if (!response.ok) {
               throw new Error(`failed to fetch remote config from ${url}: ${response.status}`)
             }
-            const wellknown = (yield* Effect.promise(() => response.json())) as { config?: Record<string, unknown> }
-            const remoteConfig = wellknown.config ?? {}
+            const wellknown = (yield* Effect.promise(() => response.json())) as {
+              config?: Record<string, unknown>
+              remote_config?: unknown
+            }
+            const remote = yield* Effect.promise(() =>
+              substituteWellKnownRemoteConfig({
+                value: wellknown.remote_config,
+                dir: url,
+                source: `${url}/.well-known/opencode`,
+              }),
+            )
+            const fetchedConfig = remote
+              ? ((yield* Effect.promise(async () => {
+                  log.debug("fetching remote config", { url: remote.url })
+                  const response = await fetch(remote.url, { headers: remote.headers })
+                  if (!response.ok)
+                    throw new Error(`failed to fetch remote config from ${remote.url}: ${response.status}`)
+                  const data = await response.json()
+                  return isRecord(data) && isRecord(data.config) ? data.config : data
+                })) as Record<string, unknown>)
+              : {}
+            const remoteConfig = mergeConfig(wellknown.config ?? {}, fetchedConfig as Info)
             if (!remoteConfig.$schema) remoteConfig.$schema = "https://bcode.sh/config.json"
             const source = `${url}/.well-known/opencode`
             const next = yield* loadConfig(JSON.stringify(remoteConfig), {
diff --git a/packages/opencode/src/control-plane/workspace.ts b/packages/opencode/src/control-plane/workspace.ts
index fe651fe3e3..f9bab469b7 100644
--- a/packages/opencode/src/control-plane/workspace.ts
+++ b/packages/opencode/src/control-plane/workspace.ts
@@ -18,18 +18,22 @@ import { ProjectID } from "@/project/schema"
 import { Slug } from "@opencode-ai/core/util/slug"
 import { WorkspaceTable } from "./workspace.sql"
 import { getAdapter } from "./adapters"
-import { type WorkspaceInfo, WorkspaceInfo as WorkspaceInfoSchema } from "./types"
+import { type Target, type WorkspaceInfo, WorkspaceInfo as WorkspaceInfoSchema } from "./types"
 import { WorkspaceID } from "./schema"
 import { Session } from "@/session/session"
 import { SessionPrompt } from "@/session/prompt"
 import { SessionTable } from "@/session/session.sql"
 import { SessionID } from "@/session/schema"
+import { NotFoundError } from "@/storage/storage"
 import { errorData } from "@/util/error"
 import { waitEvent } from "./util"
 import { WorkspaceContext } from "./workspace-context"
 import { EffectBridge } from "@/effect/bridge"
-import { NonNegativeInt, withStatics } from "@/util/schema"
+import { withStatics } from "@/util/schema"
 import { zod as effectZod, zodObject } from "@/util/effect-zod"
+import { Vcs } from "@/project/vcs"
+import { InstanceStore } from "@/project/instance-store"
+import { InstanceBootstrap } from "@/project/bootstrap"
 
 export const Info = WorkspaceInfoSchema
 export type Info = WorkspaceInfo
@@ -85,6 +89,7 @@ export type CreateInput = Schema.Schema.Type<typeof CreateInput>
 export const SessionWarpInput = Schema.Struct({
   workspaceID: Schema.NullOr(WorkspaceID),
   sessionID: SessionID,
+  copyChanges: Schema.optional(Schema.Boolean),
 }).pipe(withStatics((s) => ({ zod: effectZod(s), zodObject: zodObject(s) })))
 export type SessionWarpInput = Schema.Schema.Type<typeof SessionWarpInput>
 
@@ -136,6 +141,7 @@ type SessionWarpError =
   | WorkspaceNotFoundError
   | SessionEventsNotFoundError
   | SessionWarpHttpError
+  | Vcs.PatchApplyError
   | HttpClientError.HttpClientError
 type WaitForSyncError = SyncTimeoutError | SyncAbortedError
 type SyncLoopError = SyncHttpError | HttpClientError.HttpClientError
@@ -166,6 +172,7 @@ export const layer = Layer.effect(
     const prompt = yield* SessionPrompt.Service
     const http = yield* HttpClient.HttpClient
     const sync = yield* SyncEvent.Service
+    const vcs = yield* Vcs.Service
     const connections = new Map<WorkspaceID, ConnectionStatus>()
     const syncFibers = yield* FiberMap.make<WorkspaceID, void, SyncLoopError>()
 
@@ -254,6 +261,66 @@ export const layer = Layer.effect(
       )
     })
 
+    const runInWorkspace = <A, E, R>(input: {
+      workspaceID?: WorkspaceID
+      local: () => Effect.Effect<A, E, R>
+      remote: (input: {
+        workspace: Info
+        target: Extract<Target, { type: "remote" }>
+      }) => HttpClientRequest.HttpClientRequest
+      fallback: A
+      response?: "json" | "text"
+    }) =>
+      Effect.gen(function* () {
+        if (!input.workspaceID) return yield* input.local()
+
+        const workspace = yield* get(input.workspaceID)
+        if (!workspace) return input.fallback
+
+        const adapter = getAdapter(workspace.projectID, workspace.type)
+        const target = yield* EffectBridge.fromPromise(() => adapter.target(workspace))
+
+        if (target.type === "local") {
+          const store = yield* InstanceStore.Service
+          return yield* store.provide({ directory: target.directory }, input.local())
+        }
+
+        const response = yield* http.execute(input.remote({ workspace, target })).pipe(
+          Effect.catch((error) =>
+            Effect.sync(() => {
+              log.warn("workspace target request failed", {
+                workspaceID: workspace.id,
+                error: errorData(error),
+              })
+            }),
+          ),
+        )
+        if (!response) return input.fallback
+        if (response.status < 200 || response.status >= 300) {
+          const body = yield* response.text.pipe(Effect.catch(() => Effect.succeed("")))
+          log.warn("workspace target request failed", {
+            workspaceID: workspace.id,
+            status: response.status,
+            body,
+          })
+          return input.fallback
+        }
+
+        const body = input.response === "text" ? response.text : response.json
+        return yield* body.pipe(
+          Effect.map((result) => result as A),
+          Effect.catch((error) =>
+            Effect.sync(() => {
+              log.warn("workspace target response decode failed", {
+                workspaceID: workspace.id,
+                error: errorData(error),
+              })
+              return input.fallback
+            }),
+          ),
+        )
+      })
+
     const syncHistory = Effect.fn("Workspace.syncHistory")(function* (
       space: Info,
       url: URL | string,
@@ -556,6 +623,36 @@ export const layer = Layer.effect(
           }
         }
 
+        const sourcePatch =
+          input.copyChanges && current?.workspaceID
+            ? yield* runInWorkspace({
+                workspaceID: current?.workspaceID ?? undefined,
+                local: () => vcs.diffRaw(),
+                remote: ({ target }) =>
+                  HttpClientRequest.get(route(target.url, "/vcs/diff/raw"), {
+                    headers: new Headers(target.headers),
+                  }),
+                fallback: "",
+                response: "text",
+              }).pipe(Effect.provide(InstanceStore.defaultLayer.pipe(Layer.provide(InstanceBootstrap.defaultLayer))))
+            : ""
+
+        if (sourcePatch) {
+          // Attempt to apply the file changes to the new workspace.
+          // We intentionally do first so if it fails we don't warp
+          // the session.
+          yield* runInWorkspace({
+            workspaceID: input.workspaceID ?? undefined,
+            local: () => vcs.apply({ patch: sourcePatch }),
+            remote: ({ target }) =>
+              HttpClientRequest.post(route(target.url, "/vcs/apply"), {
+                headers: new Headers(target.headers),
+                body: HttpBody.jsonUnsafe({ patch: sourcePatch }),
+              }),
+            fallback: { applied: false },
+          }).pipe(Effect.provide(InstanceStore.defaultLayer.pipe(Layer.provide(InstanceBootstrap.defaultLayer))))
+        }
+
         if (input.workspaceID === null) {
           yield* Effect.sync(() =>
             SyncEvent.run(Session.Event.Updated, {
@@ -739,9 +836,19 @@ export const layer = Layer.effect(
 
     const remove = Effect.fn("Workspace.remove")(function* (id: WorkspaceID) {
       const sessions = yield* db((db) =>
-        db.select({ id: SessionTable.id }).from(SessionTable).where(eq(SessionTable.workspace_id, id)).all(),
+        db
+          .select({ id: SessionTable.id, parentID: SessionTable.parent_id })
+          .from(SessionTable)
+          .where(eq(SessionTable.workspace_id, id))
+          .all(),
+      )
+      const sessionIDs = new Set(sessions.map((sessionInfo) => sessionInfo.id))
+      yield* Effect.forEach(
+        sessions.filter((sessionInfo) => !sessionInfo.parentID || !sessionIDs.has(sessionInfo.parentID)),
+        (sessionInfo) =>
+          session.remove(sessionInfo.id).pipe(Effect.catchIf(NotFoundError.isInstance, () => Effect.void)),
+        { discard: true },
       )
-      yield* Effect.forEach(sessions, (sessionInfo) => session.remove(sessionInfo.id), { discard: true })
 
       const row = yield* db((db) => db.select().from(WorkspaceTable).where(eq(WorkspaceTable.id, id)).get())
       if (!row) return
@@ -855,6 +962,8 @@ export const defaultLayer = layer.pipe(
   Layer.provide(Session.defaultLayer),
   Layer.provide(SyncEvent.defaultLayer),
   Layer.provide(SessionPrompt.defaultLayer),
+  Layer.provide(Project.defaultLayer),
+  Layer.provide(Vcs.defaultLayer),
   Layer.provide(FetchHttpClient.layer),
 )
 
diff --git a/packages/opencode/src/format/index.ts b/packages/opencode/src/format/index.ts
index 7c122e3501..a61eb7be29 100644
--- a/packages/opencode/src/format/index.ts
+++ b/packages/opencode/src/format/index.ts
@@ -91,6 +91,9 @@ export const layer = Layer.effect(
                     cwd: dir,
                     env: item.environment,
                     extendEnv: true,
+                    stdin: "ignore",
+                    stdout: "ignore",
+                    stderr: "ignore",
                   }),
                 )
                 .pipe(
diff --git a/packages/opencode/src/git/index.ts b/packages/opencode/src/git/index.ts
index fff1d70b2a..349bbad466 100644
--- a/packages/opencode/src/git/index.ts
+++ b/packages/opencode/src/git/index.ts
@@ -68,6 +68,7 @@ export interface Options {
   readonly cwd: string
   readonly env?: Record<string, string>
   readonly maxOutputBytes?: number
+  readonly stdin?: ChildProcess.CommandInput
 }
 
 export interface Interface {
@@ -85,6 +86,7 @@ export interface Interface {
   readonly patchAll: (cwd: string, ref: string, options?: PatchOptions) => Effect.Effect<Patch>
   readonly patchUntracked: (cwd: string, file: string, options?: PatchOptions) => Effect.Effect<Patch>
   readonly statUntracked: (cwd: string, file: string) => Effect.Effect<Stat | undefined>
+  readonly applyPatch: (cwd: string, patch: string) => Effect.Effect<Result>
 }
 
 const kind = (code: string): Kind => {
@@ -101,6 +103,8 @@ export const layer = Layer.effect(
   Service,
   Effect.gen(function* () {
     const spawner = yield* ChildProcessSpawner.ChildProcessSpawner
+    const encoder = new TextEncoder()
+    const stdin = (text: string) => Stream.make(encoder.encode(text))
 
     const run = Effect.fn("Git.run")(
       function* (args: string[], opts: Options) {
@@ -108,7 +112,7 @@ export const layer = Layer.effect(
           cwd: opts.cwd,
           env: opts.env,
           extendEnv: true,
-          stdin: "ignore",
+          stdin: opts.stdin ?? "ignore",
           stdout: "pipe",
           stderr: "pipe",
         })
@@ -316,9 +320,13 @@ export const layer = Layer.effect(
         cwd,
         maxOutputBytes: 4096,
       })
+
       if (result.truncated) return
-      const parts = result.text().split("\t")
+      const text = result.text()
+
+      const parts = text.split("\t")
       if (parts.length < 2) return
+
       const additions = parts[0] === "-" ? 0 : Number.parseInt(parts[0] || "0", 10)
       const deletions = parts[1] === "-" ? 0 : Number.parseInt(parts[1] || "0", 10)
       return {
@@ -328,6 +336,10 @@ export const layer = Layer.effect(
       } satisfies Stat
     })
 
+    const applyPatch = Effect.fn("Git.applyPatch")(function* (cwd: string, patch: string) {
+      return yield* run(["apply", "-"], { cwd, stdin: stdin(patch) })
+    })
+
     return Service.of({
       run,
       branch,
@@ -343,6 +355,7 @@ export const layer = Layer.effect(
       patchAll,
       patchUntracked,
       statUntracked,
+      applyPatch,
     })
   }),
 )
diff --git a/packages/opencode/src/project/vcs.ts b/packages/opencode/src/project/vcs.ts
index 8b3bedbf5b..02173453db 100644
--- a/packages/opencode/src/project/vcs.ts
+++ b/packages/opencode/src/project/vcs.ts
@@ -6,7 +6,7 @@ import { InstanceState } from "@/effect/instance-state"
 import { FileWatcher } from "@/file/watcher"
 import { Git } from "@/git"
 import * as Log from "@opencode-ai/core/util/log"
-import { zod } from "@/util/effect-zod"
+import { zod, zodObject } from "@/util/effect-zod"
 import { NonNegativeInt, withStatics } from "@/util/schema"
 
 const log = Log.create({ service: "vcs" })
@@ -239,11 +239,39 @@ export const FileDiff = Schema.Struct({
   .pipe(withStatics((s) => ({ zod: zod(s) })))
 export type FileDiff = Schema.Schema.Type<typeof FileDiff>
 
+export const FileStatus = Schema.Struct({
+  file: Schema.String,
+  additions: NonNegativeInt,
+  deletions: NonNegativeInt,
+  status: Schema.Literals(["added", "deleted", "modified"]),
+})
+  .annotate({ identifier: "VcsFileStatus" })
+  .pipe(withStatics((s) => ({ zod: zod(s) })))
+export type FileStatus = Schema.Schema.Type<typeof FileStatus>
+
+export const ApplyInput = Schema.Struct({
+  patch: Schema.String,
+}).pipe(withStatics((s) => ({ zod: zod(s), zodObject: zodObject(s) })))
+export type ApplyInput = Schema.Schema.Type<typeof ApplyInput>
+
+export const ApplyResult = Schema.Struct({
+  applied: Schema.Boolean,
+}).pipe(withStatics((s) => ({ zod: zod(s) })))
+export type ApplyResult = Schema.Schema.Type<typeof ApplyResult>
+
+export class PatchApplyError extends Schema.TaggedErrorClass<PatchApplyError>()("VcsPatchApplyError", {
+  message: Schema.String,
+  reason: Schema.Literals(["non-git", "not-clean"]),
+}) {}
+
 export interface Interface {
   readonly init: () => Effect.Effect<void>
   readonly branch: () => Effect.Effect<string | undefined>
   readonly defaultBranch: () => Effect.Effect<string | undefined>
+  readonly status: () => Effect.Effect<FileStatus[]>
   readonly diff: (mode: Mode) => Effect.Effect<FileDiff[]>
+  readonly diffRaw: () => Effect.Effect<string>
+  readonly apply: (input: ApplyInput) => Effect.Effect<ApplyResult, PatchApplyError>
 }
 
 interface State {
@@ -304,6 +332,31 @@ export const layer: Layer.Layer<Service, never, Git.Service | Bus.Service> = Lay
       defaultBranch: Effect.fn("Vcs.defaultBranch")(function* () {
         return yield* InstanceState.use(state, (x) => x.root?.name)
       }),
+      status: Effect.fn("Vcs.status")(function* () {
+        const ctx = yield* InstanceState.context
+        if (ctx.project.vcs !== "git") return []
+        const ref = (yield* git.hasHead(ctx.directory)) ? "HEAD" : undefined
+        const [list, stats] = yield* Effect.all(
+          [git.status(ctx.directory), ref ? git.stats(ctx.directory, ref) : Effect.succeed([])],
+          { concurrency: 2 },
+        )
+        const map = nums(stats)
+        return yield* Effect.forEach(
+          list.toSorted((a, b) => a.file.localeCompare(b.file)),
+          (item) =>
+            Effect.gen(function* () {
+              const stat =
+                map.get(item.file) ??
+                (item.status === "added" ? yield* git.statUntracked(ctx.worktree, item.file) : undefined)
+              return {
+                file: item.file,
+                additions: stat?.additions ?? 0,
+                deletions: stat?.deletions ?? 0,
+                status: item.status,
+              } satisfies FileStatus
+            }),
+        )
+      }),
       diff: Effect.fn("Vcs.diff")(function* (mode: Mode) {
         const value = yield* InstanceState.get(state)
         const ctx = yield* InstanceState.context
@@ -318,6 +371,36 @@ export const layer: Layer.Layer<Service, never, Git.Service | Bus.Service> = Lay
         if (!ref) return []
         return yield* diffAgainstRef(git, ctx.directory, ref)
       }),
+      diffRaw: Effect.fn("Vcs.diffRaw")(function* () {
+        const ctx = yield* InstanceState.context
+        if (ctx.project.vcs !== "git") return ""
+        const [hasHead, status] = yield* Effect.all([git.hasHead(ctx.directory), git.status(ctx.directory)], {
+          concurrency: 2,
+        })
+        const tracked = hasHead ? (yield* git.patchAll(ctx.directory, "HEAD")).text : ""
+        const untracked = yield* Effect.forEach(
+          status.filter((item) => item.code === "??"),
+          (item) => git.patchUntracked(ctx.directory, item.file).pipe(Effect.map((patch) => patch.text)),
+        )
+        return [tracked, ...untracked].filter(Boolean).join("\n")
+      }),
+      apply: Effect.fn("Vcs.apply")(function* (input: ApplyInput) {
+        const ctx = yield* InstanceState.context
+        if (ctx.project.vcs !== "git") {
+          return yield* new PatchApplyError({
+            message: "Patch can't be applied because the project is not git-based",
+            reason: "non-git",
+          })
+        }
+        const applied = yield* git.applyPatch(ctx.directory, input.patch)
+        if (applied.exitCode !== 0) {
+          return yield* new PatchApplyError({
+            message: "Patch can't be applied",
+            reason: "not-clean",
+          })
+        }
+        return { applied: true }
+      }),
     })
   }),
 )
diff --git a/packages/opencode/src/provider/error.ts b/packages/opencode/src/provider/error.ts
index 3877dcb7f3..7363b5ce59 100644
--- a/packages/opencode/src/provider/error.ts
+++ b/packages/opencode/src/provider/error.ts
@@ -151,6 +151,7 @@ export function parseStreamError(input: unknown): ParsedStreamError | undefined
         isRetryable: false,
         responseBody,
       }
+    case "server_is_overloaded":
     case "server_error":
       return {
         type: "api_error",
diff --git a/packages/opencode/src/provider/transform.ts b/packages/opencode/src/provider/transform.ts
index 2fa7649c75..cd29e40822 100644
--- a/packages/opencode/src/provider/transform.ts
+++ b/packages/opencode/src/provider/transform.ts
@@ -1,4 +1,4 @@
-import type { ModelMessage } from "ai"
+import type { ModelMessage, ToolResultPart } from "ai"
 import { mergeDeep, unique } from "remeda"
 import type { JSONSchema7 } from "@ai-sdk/provider"
 import type { JSONSchema } from "zod/v4/core"
@@ -19,6 +19,10 @@ function mimeToModality(mime: string): Modality | undefined {
 
 export const OUTPUT_TOKEN_MAX = Flag.OPENCODE_EXPERIMENTAL_OUTPUT_TOKEN_MAX || 32_000
 
+export function sanitizeSurrogates(content: string) {
+  return content.replace(/[\uD800-\uDBFF](?![\uDC00-\uDFFF])|(?<![\uD800-\uDBFF])[\uDC00-\uDFFF]/g, "\uFFFD")
+}
+
 // Maps npm package to the key the AI SDK expects for providerOptions
 function sdkKey(npm: string): string | undefined {
   switch (npm) {
@@ -41,15 +45,85 @@ function sdkKey(npm: string): string | undefined {
       return "gateway"
     case "@openrouter/ai-sdk-provider":
       return "openrouter"
+    case "ai-gateway-provider":
+      // ai-gateway-provider/unified wraps createOpenAICompatible({ name: "Unified" }),
+      // and @ai-sdk/openai-compatible parses compatibleOptions from one of
+      // "openai-compatible" / "openaiCompatible" / "Unified" / "unified". The
+      // "openai-compatible" key emits a deprecation warning at runtime, so we
+      // pick the camelCase form the SDK now treats as canonical.
+      return "openaiCompatible"
   }
   return undefined
 }
 
+// TODO: fix this stupid inefficient dogshit function
 function normalizeMessages(
   msgs: ModelMessage[],
   model: Provider.Model,
   _options: Record<string, unknown>,
 ): ModelMessage[] {
+  const sanitizeToolResultOutput = (content: ToolResultPart) => {
+    if (content.output.type === "text" || content.output.type === "error-text") {
+      content.output.value = sanitizeSurrogates(content.output.value)
+    }
+    if (content.output.type === "content") {
+      content.output.value = content.output.value.map((item) => {
+        if (item.type === "text") {
+          item.text = sanitizeSurrogates(item.text)
+        }
+        return item
+      })
+    }
+    return content
+  }
+
+  msgs = msgs.map((msg) => {
+    switch (msg.role) {
+      case "tool":
+        if (!Array.isArray(msg.content)) return msg
+        msg.content = msg.content.map((content) => {
+          if (content.type === "tool-result") {
+            return sanitizeToolResultOutput(content)
+          }
+          return content
+        })
+        return msg
+
+      case "system":
+        msg.content = sanitizeSurrogates(msg.content)
+        return msg
+
+      case "user":
+        if (typeof msg.content === "string") {
+          msg.content = sanitizeSurrogates(msg.content)
+        } else {
+          msg.content = msg.content.map((content) => {
+            if (content.type === "text") {
+              content.text = sanitizeSurrogates(content.text)
+            }
+            return content
+          })
+        }
+        return msg
+
+      case "assistant":
+        if (typeof msg.content === "string") {
+          msg.content = sanitizeSurrogates(msg.content)
+        } else {
+          msg.content = msg.content.map((content) => {
+            if (content.type === "text" || content.type === "reasoning") {
+              content.text = sanitizeSurrogates(content.text)
+            }
+            if (content.type === "tool-result") {
+              return sanitizeToolResultOutput(content)
+            }
+            return content
+          })
+        }
+        return msg
+    }
+  })
+
   // Anthropic rejects messages with empty content - filter out empty string messages
   // and remove empty text/reasoning parts from array content
   if (model.api.npm === "@ai-sdk/anthropic") {
@@ -427,6 +501,36 @@ export function topK(model: Provider.Model) {
 const WIDELY_SUPPORTED_EFFORTS = ["low", "medium", "high"]
 const OPENAI_EFFORTS = ["none", "minimal", ...WIDELY_SUPPORTED_EFFORTS, "xhigh"]
 
+// OpenAI rolled out the `none` reasoning_effort tier on this date (Responses API).
+// Models released before it 400 on `reasoning_effort: "none"`, so we only expose
+// it as a variant for models new enough to accept it.
+const OPENAI_NONE_EFFORT_RELEASE_DATE = "2025-11-13"
+
+// OpenAI rolled out the `xhigh` reasoning_effort tier on this date. Same reasoning.
+const OPENAI_XHIGH_EFFORT_RELEASE_DATE = "2025-12-04"
+
+// Matches members of the gpt-5 family across the id formats we encounter:
+//   "gpt-5", "gpt-5-nano", "gpt-5.4", "openai/gpt-5.4-codex".
+// Anchored to start-of-string or "/" so it doesn't false-match "gpt-50" or "gpt-5o".
+const GPT5_FAMILY_RE = /(?:^|\/)gpt-5(?:[.-]|$)/
+
+// Computes the reasoning_effort tiers an OpenAI (or OpenAI-compatible upstream
+// routed through it, e.g. cf-ai-gateway) model exposes. Returns null for models
+// with no tunable effort knob (gpt-5-pro). Effort order: weakest to strongest.
+function openaiReasoningEfforts(apiId: string, releaseDate: string): string[] | null {
+  const id = apiId.toLowerCase()
+  if (id === "gpt-5-pro" || id === "openai/gpt-5-pro") return null
+  if (id.includes("codex")) {
+    if (id.includes("5.2") || id.includes("5.3")) return [...WIDELY_SUPPORTED_EFFORTS, "xhigh"]
+    return [...WIDELY_SUPPORTED_EFFORTS]
+  }
+  const efforts = [...WIDELY_SUPPORTED_EFFORTS]
+  if (GPT5_FAMILY_RE.test(id)) efforts.unshift("minimal")
+  if (releaseDate >= OPENAI_NONE_EFFORT_RELEASE_DATE) efforts.unshift("none")
+  if (releaseDate >= OPENAI_XHIGH_EFFORT_RELEASE_DATE) efforts.push("xhigh")
+  return efforts
+}
+
 function anthropicAdaptiveEfforts(apiId: string): string[] | null {
   if (["opus-4-7", "opus-4.7"].some((v) => apiId.includes(v))) {
     return ["low", "medium", "high", "xhigh", "max"]
@@ -476,6 +580,21 @@ export function variants(model: Provider.Model): Record<string, Record<string, a
       if (!model.id.includes("gpt") && !model.id.includes("gemini-3") && !model.id.includes("claude")) return {}
       return Object.fromEntries(OPENAI_EFFORTS.map((effort) => [effort, { reasoning: { effort } }]))
 
+    case "ai-gateway-provider": {
+      // Cloudflare AI Gateway routes every upstream through its OpenAI-compatible
+      // /v1/compat endpoint, so the body is always OAI-shaped. The gateway
+      // translates `reasoning_effort` to the upstream provider's native control
+      // (e.g. Anthropic thinking budgets) when needed. Variants therefore stay
+      // OAI-style for all upstreams, with an extended effort set for OpenAI
+      // models that support it.
+      if (model.api.id.startsWith("openai/")) {
+        const efforts = openaiReasoningEfforts(model.api.id, model.release_date)
+        if (!efforts) return {}
+        return Object.fromEntries(efforts.map((effort) => [effort, { reasoningEffort: effort }]))
+      }
+      return Object.fromEntries(WIDELY_SUPPORTED_EFFORTS.map((effort) => [effort, { reasoningEffort: effort }]))
+    }
+
     case "@ai-sdk/gateway":
       if (model.id.includes("anthropic")) {
         if (adaptiveEfforts) {
@@ -595,28 +714,12 @@ export function variants(model: Provider.Model): Record<string, Record<string, a
           },
         ]),
       )
-    case "@ai-sdk/openai":
+    case "@ai-sdk/openai": {
       // https://v5.ai-sdk.dev/providers/ai-sdk-providers/openai
-      if (id === "gpt-5-pro") return {}
-      const openaiEfforts = iife(() => {
-        if (id.includes("codex")) {
-          if (id.includes("5.2") || id.includes("5.3")) return [...WIDELY_SUPPORTED_EFFORTS, "xhigh"]
-          return WIDELY_SUPPORTED_EFFORTS
-        }
-        const arr = [...WIDELY_SUPPORTED_EFFORTS]
-        if (id.includes("gpt-5-") || id === "gpt-5") {
-          arr.unshift("minimal")
-        }
-        if (model.release_date >= "2025-11-13") {
-          arr.unshift("none")
-        }
-        if (model.release_date >= "2025-12-04") {
-          arr.push("xhigh")
-        }
-        return arr
-      })
+      const efforts = openaiReasoningEfforts(model.api.id, model.release_date)
+      if (!efforts) return {}
       return Object.fromEntries(
-        openaiEfforts.map((effort) => [
+        efforts.map((effort) => [
           effort,
           {
             reasoningEffort: effort,
@@ -625,6 +728,7 @@ export function variants(model: Provider.Model): Record<string, Record<string, a
           },
         ]),
       )
+    }
 
     case "@ai-sdk/anthropic":
     // https://v5.ai-sdk.dev/providers/ai-sdk-providers/anthropic
@@ -761,7 +865,12 @@ export function variants(model: Provider.Model): Record<string, Record<string, a
       // https://docs.mistral.ai/capabilities/reasoning/adjustable
       if (!model.capabilities.reasoning) return {}
       // Only Mistral Small 4 and Medium 3.5 support reasoning
-      const MISTRAL_REASONING_IDS = ["mistral-small-2603", "mistral-small-latest", "mistral-medium-3.5"]
+      const MISTRAL_REASONING_IDS = [
+        "mistral-small-2603",
+        "mistral-small-latest",
+        "mistral-medium-3.5",
+        "mistral-medium-2604",
+      ]
       const mistralId = model.api.id.toLowerCase()
       if (!MISTRAL_REASONING_IDS.some((id) => mistralId.includes(id))) return {}
       return {
diff --git a/packages/opencode/src/server/routes/control/workspace.ts b/packages/opencode/src/server/routes/control/workspace.ts
index 788aef3176..0c1bf252ed 100644
--- a/packages/opencode/src/server/routes/control/workspace.ts
+++ b/packages/opencode/src/server/routes/control/workspace.ts
@@ -8,6 +8,7 @@ import { AppRuntime } from "@/effect/app-runtime"
 import { WorkspaceAdapterEntry } from "@/control-plane/types"
 import { zodObject } from "@/util/effect-zod"
 import { Instance } from "@/project/instance"
+import { Vcs } from "@/project/vcs"
 import { errors } from "../../error"
 import { lazy } from "@/util/lazy"
 
@@ -164,19 +165,47 @@ export const WorkspaceRoutes = lazy(() =>
         z.object({
           id: zodObject(Workspace.Info).shape.id.nullable(),
           sessionID: Workspace.SessionWarpInput.zodObject.shape.sessionID,
+          copyChanges: z.boolean().optional(),
         }),
       ),
       async (c) => {
         const body = c.req.valid("json")
-        await AppRuntime.runPromise(
+        return AppRuntime.runPromise(
           Workspace.Service.use((workspace) =>
             workspace.sessionWarp({
               workspaceID: body.id,
               sessionID: body.sessionID,
+              copyChanges: body.copyChanges,
+            }),
+          ).pipe(
+            Effect.match({
+              onFailure: (error) => {
+                if (error instanceof Vcs.PatchApplyError) {
+                  return c.json(
+                    {
+                      name: "VcsApplyError",
+                      data: {
+                        message: error.message,
+                        reason: error.reason,
+                      },
+                    },
+                    400,
+                  )
+                }
+                return c.json(
+                  {
+                    name: "WorkspaceWarpError",
+                    data: {
+                      message: error.message,
+                    },
+                  },
+                  400,
+                )
+              },
+              onSuccess: () => c.body(null, 204),
             }),
           ),
         )
-        return c.body(null, 204)
       },
     ),
 )
diff --git a/packages/opencode/src/server/routes/instance/httpapi/AGENTS.md b/packages/opencode/src/server/routes/instance/httpapi/AGENTS.md
index 757d7aed0c..a6ccf794dd 100644
--- a/packages/opencode/src/server/routes/instance/httpapi/AGENTS.md
+++ b/packages/opencode/src/server/routes/instance/httpapi/AGENTS.md
@@ -32,4 +32,6 @@ Avoid `HttpRouter.provideRequest(...)` unless the dependency is intentionally re
 
 Use `Effect.provideService(...)` in middleware only for request-derived context, such as `WorkspaceRouteContext`, `InstanceRef`, or `WorkspaceRef`. Do not use it to smuggle stable services through request effects when they can be yielded at layer construction.
 
+Public JSON errors should be explicit `Schema.ErrorClass` contracts declared on each endpoint. Use built-in `HttpApiError.*` classes only when their empty/tagged body is the intended wire shape; for SDK-visible errors with messages, define an API error schema such as `ApiNotFoundError` and fail with that exact declared error. Keep domain and storage services free of HttpApi types, and translate expected domain errors at the handler boundary.
+
 When adding middleware, compose it at the layer boundary and keep the route tree explicit in `server.ts`. Shared router middleware such as auth, workspace routing, and instance context should stay visible where routes are assembled.
diff --git a/packages/opencode/src/server/routes/instance/httpapi/errors.ts b/packages/opencode/src/server/routes/instance/httpapi/errors.ts
new file mode 100644
index 0000000000..e5df6f5abf
--- /dev/null
+++ b/packages/opencode/src/server/routes/instance/httpapi/errors.ts
@@ -0,0 +1,18 @@
+import { Schema } from "effect"
+
+export class ApiNotFoundError extends Schema.ErrorClass<ApiNotFoundError>("NotFoundError")(
+  {
+    name: Schema.Literal("NotFoundError"),
+    data: Schema.Struct({
+      message: Schema.String,
+    }),
+  },
+  { httpApiStatus: 404 },
+) {}
+
+export function notFound(message: string) {
+  return new ApiNotFoundError({
+    name: "NotFoundError",
+    data: { message },
+  })
+}
diff --git a/packages/opencode/src/server/routes/instance/httpapi/groups/instance.ts b/packages/opencode/src/server/routes/instance/httpapi/groups/instance.ts
index 463ea1ae4c..f2b0504a05 100644
--- a/packages/opencode/src/server/routes/instance/httpapi/groups/instance.ts
+++ b/packages/opencode/src/server/routes/instance/httpapi/groups/instance.ts
@@ -5,7 +5,7 @@ import { LSP } from "@/lsp/lsp"
 import { Vcs } from "@/project/vcs"
 import { Skill } from "@/skill"
 import { Schema } from "effect"
-import { HttpApi, HttpApiEndpoint, HttpApiGroup, OpenApi } from "effect/unstable/httpapi"
+import { HttpApi, HttpApiEndpoint, HttpApiGroup, HttpApiSchema, OpenApi } from "effect/unstable/httpapi"
 import { Authorization } from "../middleware/authorization"
 import { InstanceContextMiddleware } from "../middleware/instance-context"
 import { WorkspaceRoutingMiddleware } from "../middleware/workspace-routing"
@@ -23,11 +23,25 @@ export const VcsDiffQuery = Schema.Struct({
   mode: Vcs.Mode,
 })
 
+export class ApiVcsApplyError extends Schema.ErrorClass<ApiVcsApplyError>("VcsApplyError")(
+  {
+    name: Schema.Literal("VcsApplyError"),
+    data: Schema.Struct({
+      message: Schema.String,
+      reason: Schema.Literals(["non-git", "not-clean"]),
+    }),
+  },
+  { httpApiStatus: 400 },
+) {}
+
 export const InstancePaths = {
   dispose: "/instance/dispose",
   path: "/path",
   vcs: "/vcs",
+  vcsStatus: "/vcs/status",
   vcsDiff: "/vcs/diff",
+  vcsDiffRaw: "/vcs/diff/raw",
+  vcsApply: "/vcs/apply",
   command: "/command",
   agent: "/agent",
   skill: "/skill",
@@ -68,6 +82,15 @@ export const InstanceApi = HttpApi.make("instance")
               "Retrieve version control system (VCS) information for the current project, such as git branch.",
           }),
         ),
+        HttpApiEndpoint.get("vcsStatus", InstancePaths.vcsStatus, {
+          success: described(Schema.Array(Vcs.FileStatus), "VCS status"),
+        }).annotateMerge(
+          OpenApi.annotations({
+            identifier: "vcs.status",
+            summary: "Get VCS status",
+            description: "Retrieve changed files in the current working tree without patches.",
+          }),
+        ),
         HttpApiEndpoint.get("vcsDiff", InstancePaths.vcsDiff, {
           query: VcsDiffQuery,
           success: described(Schema.Array(Vcs.FileDiff), "VCS diff"),
@@ -78,6 +101,29 @@ export const InstanceApi = HttpApi.make("instance")
             description: "Retrieve the current git diff for the working tree or against the default branch.",
           }),
         ),
+        HttpApiEndpoint.get("vcsDiffRaw", InstancePaths.vcsDiffRaw, {
+          success: described(
+            Schema.String.pipe(HttpApiSchema.asText({ contentType: "text/x-diff; charset=utf-8" })),
+            "Raw VCS diff",
+          ),
+        }).annotateMerge(
+          OpenApi.annotations({
+            identifier: "vcs.diff.raw",
+            summary: "Get raw VCS diff",
+            description: "Retrieve a raw patch for current uncommitted changes.",
+          }),
+        ),
+        HttpApiEndpoint.post("vcsApply", InstancePaths.vcsApply, {
+          payload: Vcs.ApplyInput,
+          success: described(Vcs.ApplyResult, "VCS patch applied"),
+          error: ApiVcsApplyError,
+        }).annotateMerge(
+          OpenApi.annotations({
+            identifier: "vcs.apply",
+            summary: "Apply VCS patch",
+            description: "Apply a raw patch to the current working tree.",
+          }),
+        ),
         HttpApiEndpoint.get("command", InstancePaths.command, {
           success: described(Schema.Array(Command.Info), "List of commands"),
         }).annotateMerge(
diff --git a/packages/opencode/src/server/routes/instance/httpapi/groups/pty.ts b/packages/opencode/src/server/routes/instance/httpapi/groups/pty.ts
index 3304ab9fbf..ad513e0ad4 100644
--- a/packages/opencode/src/server/routes/instance/httpapi/groups/pty.ts
+++ b/packages/opencode/src/server/routes/instance/httpapi/groups/pty.ts
@@ -6,6 +6,7 @@ import { HttpApi, HttpApiEndpoint, HttpApiError, HttpApiGroup, OpenApi } from "e
 import { Authorization } from "../middleware/authorization"
 import { InstanceContextMiddleware } from "../middleware/instance-context"
 import { WorkspaceRoutingMiddleware } from "../middleware/workspace-routing"
+import { ApiNotFoundError } from "../errors"
 import { described } from "./metadata"
 
 const root = "/pty"
@@ -64,7 +65,7 @@ export const PtyApi = HttpApi.make("pty")
         HttpApiEndpoint.get("get", PtyPaths.get, {
           params: { ptyID: PtyID },
           success: described(Pty.Info, "Session info"),
-          error: HttpApiError.NotFound,
+          error: ApiNotFoundError,
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "pty.get",
@@ -76,7 +77,7 @@ export const PtyApi = HttpApi.make("pty")
           params: { ptyID: PtyID },
           payload: Pty.UpdateInput,
           success: described(Pty.Info, "Updated session"),
-          error: [HttpApiError.BadRequest, HttpApiError.NotFound],
+          error: [HttpApiError.BadRequest, ApiNotFoundError],
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "pty.update",
@@ -87,7 +88,7 @@ export const PtyApi = HttpApi.make("pty")
         HttpApiEndpoint.delete("remove", PtyPaths.remove, {
           params: { ptyID: PtyID },
           success: described(Schema.Boolean, "Session removed"),
-          error: HttpApiError.NotFound,
+          error: ApiNotFoundError,
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "pty.remove",
@@ -98,7 +99,7 @@ export const PtyApi = HttpApi.make("pty")
         HttpApiEndpoint.post("connectToken", PtyPaths.connectToken, {
           params: { ptyID: PtyID },
           success: described(PtyTicket.ConnectToken, "WebSocket connect token"),
-          error: [HttpApiError.Forbidden, HttpApiError.NotFound],
+          error: [HttpApiError.Forbidden, ApiNotFoundError],
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "pty.connectToken",
diff --git a/packages/opencode/src/server/routes/instance/httpapi/groups/session.ts b/packages/opencode/src/server/routes/instance/httpapi/groups/session.ts
index 77d064ff5a..1159c88030 100644
--- a/packages/opencode/src/server/routes/instance/httpapi/groups/session.ts
+++ b/packages/opencode/src/server/routes/instance/httpapi/groups/session.ts
@@ -15,6 +15,7 @@ import { HttpApi, HttpApiEndpoint, HttpApiError, HttpApiGroup, HttpApiSchema, Op
 import { Authorization } from "../middleware/authorization"
 import { InstanceContextMiddleware } from "../middleware/instance-context"
 import { WorkspaceRoutingMiddleware } from "../middleware/workspace-routing"
+import { ApiNotFoundError } from "../errors"
 import { described } from "./metadata"
 
 const root = "/session"
@@ -123,7 +124,7 @@ export const SessionApi = HttpApi.make("session")
         HttpApiEndpoint.get("get", SessionPaths.get, {
           params: { sessionID: SessionID },
           success: described(Session.Info, "Get session"),
-          error: [HttpApiError.BadRequest, HttpApiError.NotFound],
+          error: [HttpApiError.BadRequest, ApiNotFoundError],
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "session.get",
@@ -168,7 +169,7 @@ export const SessionApi = HttpApi.make("session")
           params: { sessionID: SessionID },
           query: MessagesQuery,
           success: described(Schema.Array(MessageV2.WithParts), "List of messages"),
-          error: [HttpApiError.BadRequest, HttpApiError.NotFound],
+          error: [HttpApiError.BadRequest, ApiNotFoundError],
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "session.messages",
@@ -179,7 +180,7 @@ export const SessionApi = HttpApi.make("session")
         HttpApiEndpoint.get("message", SessionPaths.message, {
           params: { sessionID: SessionID, messageID: MessageID },
           success: described(MessageV2.WithParts, "Message"),
-          error: [HttpApiError.BadRequest, HttpApiError.NotFound],
+          error: [HttpApiError.BadRequest, ApiNotFoundError],
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "session.message",
@@ -201,7 +202,7 @@ export const SessionApi = HttpApi.make("session")
         HttpApiEndpoint.delete("remove", SessionPaths.remove, {
           params: { sessionID: SessionID },
           success: described(Schema.Boolean, "Successfully deleted session"),
-          error: [HttpApiError.BadRequest, HttpApiError.NotFound],
+          error: [HttpApiError.BadRequest, ApiNotFoundError],
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "session.delete",
@@ -213,7 +214,7 @@ export const SessionApi = HttpApi.make("session")
           params: { sessionID: SessionID },
           payload: UpdatePayload,
           success: described(Session.Info, "Successfully updated session"),
-          error: [HttpApiError.BadRequest, HttpApiError.NotFound],
+          error: [HttpApiError.BadRequest, ApiNotFoundError],
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "session.update",
@@ -225,6 +226,7 @@ export const SessionApi = HttpApi.make("session")
           params: { sessionID: SessionID },
           payload: ForkPayload,
           success: described(Session.Info, "200"),
+          error: ApiNotFoundError,
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "session.fork",
@@ -259,7 +261,7 @@ export const SessionApi = HttpApi.make("session")
         HttpApiEndpoint.post("share", SessionPaths.share, {
           params: { sessionID: SessionID },
           success: described(Session.Info, "Successfully shared session"),
-          error: [HttpApiError.BadRequest, HttpApiError.NotFound],
+          error: [HttpApiError.BadRequest, ApiNotFoundError],
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "session.share",
@@ -270,7 +272,7 @@ export const SessionApi = HttpApi.make("session")
         HttpApiEndpoint.delete("unshare", SessionPaths.share, {
           params: { sessionID: SessionID },
           success: described(Session.Info, "Successfully unshared session"),
-          error: [HttpApiError.BadRequest, HttpApiError.NotFound],
+          error: [HttpApiError.BadRequest, ApiNotFoundError],
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "session.unshare",
@@ -282,7 +284,7 @@ export const SessionApi = HttpApi.make("session")
           params: { sessionID: SessionID },
           payload: SummarizePayload,
           success: described(Schema.Boolean, "Summarized session"),
-          error: [HttpApiError.BadRequest, HttpApiError.NotFound],
+          error: [HttpApiError.BadRequest, ApiNotFoundError],
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "session.summarize",
diff --git a/packages/opencode/src/server/routes/instance/httpapi/groups/tui.ts b/packages/opencode/src/server/routes/instance/httpapi/groups/tui.ts
index efe73d95d1..8ab43f6654 100644
--- a/packages/opencode/src/server/routes/instance/httpapi/groups/tui.ts
+++ b/packages/opencode/src/server/routes/instance/httpapi/groups/tui.ts
@@ -4,6 +4,7 @@ import { HttpApi, HttpApiEndpoint, HttpApiError, HttpApiGroup, OpenApi } from "e
 import { Authorization } from "../middleware/authorization"
 import { InstanceContextMiddleware } from "../middleware/instance-context"
 import { WorkspaceRoutingMiddleware } from "../middleware/workspace-routing"
+import { ApiNotFoundError } from "../errors"
 import { described } from "./metadata"
 
 const root = "/tui"
@@ -155,7 +156,7 @@ export const TuiApi = HttpApi.make("tui")
         HttpApiEndpoint.post("selectSession", TuiPaths.selectSession, {
           payload: TuiEvent.SessionSelect.properties,
           success: described(Schema.Boolean, "Session selected successfully"),
-          error: [HttpApiError.BadRequest, HttpApiError.NotFound],
+          error: [HttpApiError.BadRequest, ApiNotFoundError],
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "tui.selectSession",
diff --git a/packages/opencode/src/server/routes/instance/httpapi/groups/workspace.ts b/packages/opencode/src/server/routes/instance/httpapi/groups/workspace.ts
index f197ab9765..66422c13b6 100644
--- a/packages/opencode/src/server/routes/instance/httpapi/groups/workspace.ts
+++ b/packages/opencode/src/server/routes/instance/httpapi/groups/workspace.ts
@@ -2,6 +2,7 @@ import { Workspace } from "@/control-plane/workspace"
 import { WorkspaceAdapterEntry } from "@/control-plane/types"
 import { Schema, Struct } from "effect"
 import { HttpApi, HttpApiEndpoint, HttpApiError, HttpApiGroup, HttpApiSchema, OpenApi } from "effect/unstable/httpapi"
+import { ApiVcsApplyError } from "./instance"
 import { Authorization } from "../middleware/authorization"
 import { InstanceContextMiddleware } from "../middleware/instance-context"
 import { WorkspaceRoutingMiddleware } from "../middleware/workspace-routing"
@@ -12,8 +13,19 @@ export const CreatePayload = Schema.Struct(Struct.omit(Workspace.CreateInput.fie
 export const WarpPayload = Schema.Struct({
   id: Schema.NullOr(Workspace.Info.fields.id),
   sessionID: Workspace.SessionWarpInput.fields.sessionID,
+  copyChanges: Workspace.SessionWarpInput.fields.copyChanges,
 })
 
+export class ApiWorkspaceWarpError extends Schema.ErrorClass<ApiWorkspaceWarpError>("WorkspaceWarpError")(
+  {
+    name: Schema.Literal("WorkspaceWarpError"),
+    data: Schema.Struct({
+      message: Schema.String,
+    }),
+  },
+  { httpApiStatus: 400 },
+) {}
+
 export const WorkspacePaths = {
   adapters: `${root}/adapter`,
   list: root,
@@ -78,7 +90,7 @@ export const WorkspaceApi = HttpApi.make("workspace")
         HttpApiEndpoint.post("warp", WorkspacePaths.warp, {
           payload: WarpPayload,
           success: described(HttpApiSchema.NoContent, "Session warped"),
-          error: HttpApiError.BadRequest,
+          error: [ApiWorkspaceWarpError, ApiVcsApplyError],
         }).annotateMerge(
           OpenApi.annotations({
             identifier: "experimental.workspace.warp",
diff --git a/packages/opencode/src/server/routes/instance/httpapi/handlers/instance.ts b/packages/opencode/src/server/routes/instance/httpapi/handlers/instance.ts
index c2a4503b48..50a7fecfa7 100644
--- a/packages/opencode/src/server/routes/instance/httpapi/handlers/instance.ts
+++ b/packages/opencode/src/server/routes/instance/httpapi/handlers/instance.ts
@@ -9,6 +9,7 @@ import { Skill } from "@/skill"
 import { Effect } from "effect"
 import { HttpApiBuilder } from "effect/unstable/httpapi"
 import { InstanceHttpApi } from "../api"
+import { ApiVcsApplyError } from "../groups/instance"
 import { markInstanceForDisposal } from "../lifecycle"
 
 export const instanceHandlers = HttpApiBuilder.group(InstanceHttpApi, "instance", (handlers) =>
@@ -41,10 +42,33 @@ export const instanceHandlers = HttpApiBuilder.group(InstanceHttpApi, "instance"
       return { branch, default_branch }
     })
 
+    const getVcsStatus = Effect.fn("InstanceHttpApi.vcsStatus")(function* () {
+      return yield* vcs.status()
+    })
+
     const getVcsDiff = Effect.fn("InstanceHttpApi.vcsDiff")(function* (ctx: { query: { mode: Vcs.Mode } }) {
       return yield* vcs.diff(ctx.query.mode)
     })
 
+    const getVcsDiffRaw = Effect.fn("InstanceHttpApi.vcsDiffRaw")(function* () {
+      return yield* vcs.diffRaw()
+    })
+
+    const applyVcs = Effect.fn("InstanceHttpApi.vcsApply")(function* (ctx: { payload: Vcs.ApplyInput }) {
+      return yield* vcs.apply(ctx.payload).pipe(
+        Effect.mapError(
+          (error) =>
+            new ApiVcsApplyError({
+              name: "VcsApplyError",
+              data: {
+                message: error.message,
+                reason: error.reason,
+              },
+            }),
+        ),
+      )
+    })
+
     const getCommand = Effect.fn("InstanceHttpApi.command")(function* () {
       return yield* command.list()
     })
@@ -69,7 +93,10 @@ export const instanceHandlers = HttpApiBuilder.group(InstanceHttpApi, "instance"
       .handle("dispose", dispose)
       .handle("path", getPath)
       .handle("vcs", getVcs)
+      .handle("vcsStatus", getVcsStatus)
       .handle("vcsDiff", getVcsDiff)
+      .handle("vcsDiffRaw", getVcsDiffRaw)
+      .handle("vcsApply", applyVcs)
       .handle("command", getCommand)
       .handle("agent", getAgent)
       .handle("skill", getSkill)
diff --git a/packages/opencode/src/server/routes/instance/httpapi/handlers/pty.ts b/packages/opencode/src/server/routes/instance/httpapi/handlers/pty.ts
index e5ff300a2a..7b8395d809 100644
--- a/packages/opencode/src/server/routes/instance/httpapi/handlers/pty.ts
+++ b/packages/opencode/src/server/routes/instance/httpapi/handlers/pty.ts
@@ -15,6 +15,7 @@ import { HttpRouter, HttpServerRequest, HttpServerResponse } from "effect/unstab
 import { HttpApiBuilder, HttpApiError } from "effect/unstable/httpapi"
 import * as Socket from "effect/unstable/socket/Socket"
 import { InstanceHttpApi } from "../api"
+import * as ApiError from "../errors"
 import { CursorQuery, Params, PtyPaths } from "../groups/pty"
 import { WebSocketTracker } from "../websocket-tracker"
 
@@ -46,7 +47,7 @@ export const ptyHandlers = HttpApiBuilder.group(InstanceHttpApi, "pty", (handler
 
     const get = Effect.fn("PtyHttpApi.get")(function* (ctx: { params: { ptyID: PtyID } }) {
       const info = yield* pty.get(ctx.params.ptyID)
-      if (!info) return yield* new HttpApiError.NotFound({})
+      if (!info) return yield* ApiError.notFound("Session not found")
       return info
     })
 
@@ -58,7 +59,7 @@ export const ptyHandlers = HttpApiBuilder.group(InstanceHttpApi, "pty", (handler
         ...ctx.payload,
         size: ctx.payload.size ? { ...ctx.payload.size } : undefined,
       })
-      if (!info) return yield* new HttpApiError.NotFound({})
+      if (!info) return yield* ApiError.notFound("Session not found")
       return info
     })
 
@@ -71,7 +72,7 @@ export const ptyHandlers = HttpApiBuilder.group(InstanceHttpApi, "pty", (handler
       const request = yield* HttpServerRequest.HttpServerRequest
       if (request.headers[PTY_CONNECT_TOKEN_HEADER] !== PTY_CONNECT_TOKEN_HEADER_VALUE || !validOrigin(request, cors))
         return yield* new HttpApiError.Forbidden({})
-      if (!(yield* pty.get(ctx.params.ptyID))) return yield* new HttpApiError.NotFound({})
+      if (!(yield* pty.get(ctx.params.ptyID))) return yield* ApiError.notFound("Session not found")
       return yield* tickets.issue({ ptyID: ctx.params.ptyID, ...(yield* PtyTicket.scope) })
     })
 
diff --git a/packages/opencode/src/server/routes/instance/httpapi/handlers/session-errors.ts b/packages/opencode/src/server/routes/instance/httpapi/handlers/session-errors.ts
new file mode 100644
index 0000000000..98ac2b9ad6
--- /dev/null
+++ b/packages/opencode/src/server/routes/instance/httpapi/handlers/session-errors.ts
@@ -0,0 +1,9 @@
+import type { NotFoundError as StorageNotFoundError } from "@/storage/storage"
+import { Effect } from "effect"
+import * as ApiError from "../errors"
+
+type StorageNotFound = InstanceType<typeof StorageNotFoundError>
+
+export function mapStorageNotFound<A, R>(self: Effect.Effect<A, StorageNotFound, R>) {
+  return self.pipe(Effect.mapError((error) => ApiError.notFound(error.data.message)))
+}
diff --git a/packages/opencode/src/server/routes/instance/httpapi/handlers/session.ts b/packages/opencode/src/server/routes/instance/httpapi/handlers/session.ts
index 4a67ba036e..56fa7adb15 100644
--- a/packages/opencode/src/server/routes/instance/httpapi/handlers/session.ts
+++ b/packages/opencode/src/server/routes/instance/httpapi/handlers/session.ts
@@ -37,14 +37,7 @@ import {
   SummarizePayload,
   UpdatePayload,
 } from "../groups/session"
-
-const mapNotFound = <A, E, R>(self: Effect.Effect<A, E, R>) =>
-  self.pipe(
-    Effect.catchIf(NotFoundError.isInstance, () => Effect.fail(new HttpApiError.NotFound({}))),
-    Effect.catchDefect((error) =>
-      NotFoundError.isInstance(error) ? Effect.fail(new HttpApiError.NotFound({})) : Effect.die(error),
-    ),
-  )
+import * as SessionError from "./session-errors"
 
 export const sessionHandlers = HttpApiBuilder.group(InstanceHttpApi, "session", (handlers) =>
   Effect.gen(function* () {
@@ -79,7 +72,7 @@ export const sessionHandlers = HttpApiBuilder.group(InstanceHttpApi, "session",
     })
 
     const get = Effect.fn("SessionHttpApi.get")(function* (ctx: { params: { sessionID: SessionID } }) {
-      return yield* mapNotFound(session.get(ctx.params.sessionID))
+      return yield* SessionError.mapStorageNotFound(session.get(ctx.params.sessionID))
     })
 
     const children = Effect.fn("SessionHttpApi.children")(function* (ctx: { params: { sessionID: SessionID } }) {
@@ -101,51 +94,49 @@ export const sessionHandlers = HttpApiBuilder.group(InstanceHttpApi, "session",
       params: { sessionID: SessionID }
       query: typeof MessagesQuery.Type
     }) {
-      return yield* mapNotFound(
-        Effect.gen(function* () {
-          if (ctx.query.before && ctx.query.limit === undefined) return yield* new HttpApiError.BadRequest({})
-          if (ctx.query.before) {
-            const before = ctx.query.before
-            yield* Effect.try({
-              try: () => MessageV2.cursor.decode(before),
-              catch: () => new HttpApiError.BadRequest({}),
-            })
-          }
-          if (ctx.query.limit === undefined || ctx.query.limit === 0) {
-            yield* session.get(ctx.params.sessionID)
-            return yield* session.messages({ sessionID: ctx.params.sessionID })
-          }
-
-          yield* session.get(ctx.params.sessionID)
-          const page = MessageV2.page({
-            sessionID: ctx.params.sessionID,
-            limit: ctx.query.limit,
-            before: ctx.query.before,
-          })
-          if (!page.cursor) return page.items
-
-          const request = yield* HttpServerRequest.HttpServerRequest
-          // toURL() honors the Host + x-forwarded-proto headers, so the Link
-          // header echoes the real origin instead of a hard-coded localhost.
-          const url = Option.getOrElse(HttpServerRequest.toURL(request), () => new URL(request.url, "http://localhost"))
-          url.searchParams.set("limit", ctx.query.limit.toString())
-          url.searchParams.set("before", page.cursor)
-          return HttpServerResponse.jsonUnsafe(page.items, {
-            headers: {
-              "Access-Control-Expose-Headers": "Link, X-Next-Cursor",
-              Link: `<${url.toString()}>; rel="next"`,
-              "X-Next-Cursor": page.cursor,
-            },
-          })
-        }),
-      )
+      if (ctx.query.before && ctx.query.limit === undefined) return yield* new HttpApiError.BadRequest({})
+      if (ctx.query.before) {
+        const before = ctx.query.before
+        yield* Effect.try({
+          try: () => MessageV2.cursor.decode(before),
+          catch: () => new HttpApiError.BadRequest({}),
+        })
+      }
+      yield* SessionError.mapStorageNotFound(session.get(ctx.params.sessionID))
+      if (ctx.query.limit === undefined || ctx.query.limit === 0) {
+        return yield* session.messages({ sessionID: ctx.params.sessionID })
+      }
+
+      const page = MessageV2.page({
+        sessionID: ctx.params.sessionID,
+        limit: ctx.query.limit,
+        before: ctx.query.before,
+      })
+      if (!page.cursor) return page.items
+
+      const request = yield* HttpServerRequest.HttpServerRequest
+      // toURL() honors the Host + x-forwarded-proto headers, so the Link
+      // header echoes the real origin instead of a hard-coded localhost.
+      const url = Option.getOrElse(HttpServerRequest.toURL(request), () => new URL(request.url, "http://localhost"))
+      url.searchParams.set("limit", ctx.query.limit.toString())
+      url.searchParams.set("before", page.cursor)
+      return HttpServerResponse.jsonUnsafe(page.items, {
+        headers: {
+          "Access-Control-Expose-Headers": "Link, X-Next-Cursor",
+          Link: `<${url.toString()}>; rel="next"`,
+          "X-Next-Cursor": page.cursor,
+        },
+      })
     })
 
     const message = Effect.fn("SessionHttpApi.message")(function* (ctx: {
       params: { sessionID: SessionID; messageID: MessageID }
     }) {
-      return yield* mapNotFound(
-        Effect.sync(() => MessageV2.get({ sessionID: ctx.params.sessionID, messageID: ctx.params.messageID })),
+      return yield* SessionError.mapStorageNotFound(
+        Effect.try({
+          try: () => MessageV2.get({ sessionID: ctx.params.sessionID, messageID: ctx.params.messageID }),
+          catch: (error) => error,
+        }).pipe(Effect.catch((error) => (NotFoundError.isInstance(error) ? Effect.fail(error) : Effect.die(error)))),
       )
     })
 
@@ -170,7 +161,7 @@ export const sessionHandlers = HttpApiBuilder.group(InstanceHttpApi, "session",
     })
 
     const remove = Effect.fn("SessionHttpApi.remove")(function* (ctx: { params: { sessionID: SessionID } }) {
-      yield* session.remove(ctx.params.sessionID)
+      yield* SessionError.mapStorageNotFound(session.remove(ctx.params.sessionID))
       return true
     })
 
@@ -178,7 +169,7 @@ export const sessionHandlers = HttpApiBuilder.group(InstanceHttpApi, "session",
       params: { sessionID: SessionID }
       payload: typeof UpdatePayload.Type
     }) {
-      const current = yield* session.get(ctx.params.sessionID)
+      const current = yield* SessionError.mapStorageNotFound(session.get(ctx.params.sessionID))
       if (ctx.payload.title !== undefined) {
         yield* session.setTitle({ sessionID: ctx.params.sessionID, title: ctx.payload.title })
       }
@@ -191,14 +182,16 @@ export const sessionHandlers = HttpApiBuilder.group(InstanceHttpApi, "session",
       if (ctx.payload.time?.archived !== undefined) {
         yield* session.setArchived({ sessionID: ctx.params.sessionID, time: ctx.payload.time.archived })
       }
-      return yield* session.get(ctx.params.sessionID)
+      return yield* SessionError.mapStorageNotFound(session.get(ctx.params.sessionID))
     })
 
     const fork = Effect.fn("SessionHttpApi.fork")(function* (ctx: {
       params: { sessionID: SessionID }
       payload: typeof ForkPayload.Type
     }) {
-      return yield* session.fork({ sessionID: ctx.params.sessionID, messageID: ctx.payload.messageID })
+      return yield* SessionError.mapStorageNotFound(
+        session.fork({ sessionID: ctx.params.sessionID, messageID: ctx.payload.messageID }),
+      )
     })
 
     const abort = Effect.fn("SessionHttpApi.abort")(function* (ctx: { params: { sessionID: SessionID } }) {
@@ -222,19 +215,19 @@ export const sessionHandlers = HttpApiBuilder.group(InstanceHttpApi, "session",
 
     const share = Effect.fn("SessionHttpApi.share")(function* (ctx: { params: { sessionID: SessionID } }) {
       yield* shareSvc.share(ctx.params.sessionID).pipe(Effect.mapError(() => new HttpApiError.BadRequest({})))
-      return yield* session.get(ctx.params.sessionID)
+      return yield* SessionError.mapStorageNotFound(session.get(ctx.params.sessionID))
     })
 
     const unshare = Effect.fn("SessionHttpApi.unshare")(function* (ctx: { params: { sessionID: SessionID } }) {
       yield* shareSvc.unshare(ctx.params.sessionID).pipe(Effect.mapError(() => new HttpApiError.BadRequest({})))
-      return yield* session.get(ctx.params.sessionID)
+      return yield* SessionError.mapStorageNotFound(session.get(ctx.params.sessionID))
     })
 
     const summarize = Effect.fn("SessionHttpApi.summarize")(function* (ctx: {
       params: { sessionID: SessionID }
       payload: typeof SummarizePayload.Type
     }) {
-      yield* revertSvc.cleanup(yield* session.get(ctx.params.sessionID))
+      yield* revertSvc.cleanup(yield* SessionError.mapStorageNotFound(session.get(ctx.params.sessionID)))
       const messages = yield* session.messages({ sessionID: ctx.params.sessionID })
       const defaultAgent = yield* agentSvc.defaultAgent()
       const currentAgent = messages.findLast((message) => message.info.role === "user")?.info.agent ?? defaultAgent
diff --git a/packages/opencode/src/server/routes/instance/httpapi/handlers/tui.ts b/packages/opencode/src/server/routes/instance/httpapi/handlers/tui.ts
index cc85321685..0ecebf451f 100644
--- a/packages/opencode/src/server/routes/instance/httpapi/handlers/tui.ts
+++ b/packages/opencode/src/server/routes/instance/httpapi/handlers/tui.ts
@@ -1,13 +1,12 @@
 import { Bus } from "@/bus"
 import { TuiEvent } from "@/cli/cmd/tui/event"
-import { SessionTable } from "@/session/session.sql"
-import * as Database from "@/storage/db"
-import { eq } from "drizzle-orm"
+import { Session } from "@/session/session"
 import { Effect } from "effect"
 import { HttpApiBuilder, HttpApiError } from "effect/unstable/httpapi"
 import { nextTuiRequest, submitTuiResponse } from "@/server/shared/tui-control"
 import { InstanceHttpApi } from "../api"
 import { CommandPayload, TuiPublishPayload } from "../groups/tui"
+import * as SessionError from "./session-errors"
 
 const commandAliases = {
   session_new: "session.new",
@@ -28,6 +27,7 @@ const commandAliases = {
 export const tuiHandlers = HttpApiBuilder.group(InstanceHttpApi, "tui", (handlers) =>
   Effect.gen(function* () {
     const bus = yield* Bus.Service
+    const session = yield* Session.Service
     const publishCommand = (command: typeof TuiEvent.CommandExecute.properties.Type.command | undefined) =>
       bus.publish(TuiEvent.CommandExecute, { command } as typeof TuiEvent.CommandExecute.properties.Type)
 
@@ -98,12 +98,7 @@ export const tuiHandlers = HttpApiBuilder.group(InstanceHttpApi, "tui", (handler
       payload: typeof TuiEvent.SessionSelect.properties.Type
     }) {
       if (!ctx.payload.sessionID.startsWith("ses")) return yield* new HttpApiError.BadRequest({})
-      const row = yield* Effect.sync(() =>
-        Database.use((db) =>
-          db.select({ id: SessionTable.id }).from(SessionTable).where(eq(SessionTable.id, ctx.payload.sessionID)).get(),
-        ),
-      )
-      if (!row) return yield* new HttpApiError.NotFound({})
+      yield* SessionError.mapStorageNotFound(session.get(ctx.payload.sessionID))
       yield* bus.publish(TuiEvent.SessionSelect, ctx.payload)
       return true
     })
diff --git a/packages/opencode/src/server/routes/instance/httpapi/handlers/workspace.ts b/packages/opencode/src/server/routes/instance/httpapi/handlers/workspace.ts
index b415943a62..d908eda9d1 100644
--- a/packages/opencode/src/server/routes/instance/httpapi/handlers/workspace.ts
+++ b/packages/opencode/src/server/routes/instance/httpapi/handlers/workspace.ts
@@ -1,10 +1,12 @@
 import { listAdapters } from "@/control-plane/adapters"
 import { Workspace } from "@/control-plane/workspace"
 import * as InstanceState from "@/effect/instance-state"
+import { Vcs } from "@/project/vcs"
 import { Effect } from "effect"
 import { HttpApiBuilder, HttpApiError } from "effect/unstable/httpapi"
 import { InstanceHttpApi } from "../api"
-import { CreatePayload, WarpPayload } from "../groups/workspace"
+import { ApiVcsApplyError } from "../groups/instance"
+import { ApiWorkspaceWarpError, CreatePayload, WarpPayload } from "../groups/workspace"
 
 export const workspaceHandlers = HttpApiBuilder.group(InstanceHttpApi, "workspace", (handlers) =>
   Effect.gen(function* () {
@@ -44,8 +46,27 @@ export const workspaceHandlers = HttpApiBuilder.group(InstanceHttpApi, "workspac
         .sessionWarp({
           workspaceID: ctx.payload.id,
           sessionID: ctx.payload.sessionID,
+          copyChanges: ctx.payload.copyChanges,
         })
-        .pipe(Effect.mapError(() => new HttpApiError.BadRequest({})))
+        .pipe(
+          Effect.mapError((error) => {
+            if (error instanceof Vcs.PatchApplyError) {
+              return new ApiVcsApplyError({
+                name: "VcsApplyError",
+                data: {
+                  message: error.message,
+                  reason: error.reason,
+                },
+              })
+            }
+            return new ApiWorkspaceWarpError({
+              name: "WorkspaceWarpError",
+              data: {
+                message: error.message,
+              },
+            })
+          }),
+        )
     })
 
     return handlers
diff --git a/packages/opencode/src/server/routes/instance/httpapi/middleware/workspace-routing.ts b/packages/opencode/src/server/routes/instance/httpapi/middleware/workspace-routing.ts
index a91a9992df..8ec9f74860 100644
--- a/packages/opencode/src/server/routes/instance/httpapi/middleware/workspace-routing.ts
+++ b/packages/opencode/src/server/routes/instance/httpapi/middleware/workspace-routing.ts
@@ -7,6 +7,7 @@ import { Session } from "@/session/session"
 import { HttpApiProxy } from "./proxy"
 import * as Fence from "@/server/shared/fence"
 import { getWorkspaceRouteSessionID, isLocalWorkspaceRoute, workspaceProxyURL } from "@/server/shared/workspace-routing"
+import { NotFoundError } from "@/storage/storage"
 import { Flag } from "@opencode-ai/core/flag/flag"
 import { Context, Data, Effect, Layer } from "effect"
 import { HttpClient, HttpRouter, HttpServerRequest, HttpServerResponse } from "effect/unstable/http"
@@ -178,7 +179,10 @@ function routeHttpApiWorkspace<E>(
     const request = yield* HttpServerRequest.HttpServerRequest
     const sessionID = getWorkspaceRouteSessionID(requestURL(request))
     const session = sessionID
-      ? yield* Session.Service.use((svc) => svc.get(sessionID)).pipe(Effect.catchDefect(() => Effect.void))
+      ? yield* Session.Service.use((svc) => svc.get(sessionID)).pipe(
+          Effect.catchIf(NotFoundError.isInstance, () => Effect.succeed(undefined)),
+          Effect.catchDefect(() => Effect.succeed(undefined)),
+        )
       : undefined
     const plan = yield* planRequest(request, session?.workspaceID)
     return yield* routeWorkspace(client, effect, plan)
diff --git a/packages/opencode/src/server/routes/instance/httpapi/public.ts b/packages/opencode/src/server/routes/instance/httpapi/public.ts
index c9668336ae..b2ac719a2a 100644
--- a/packages/opencode/src/server/routes/instance/httpapi/public.ts
+++ b/packages/opencode/src/server/routes/instance/httpapi/public.ts
@@ -146,6 +146,16 @@ function matchLegacyOpenApi(input: Record<string, unknown>) {
           if (properties?.branch) properties.branch = { anyOf: [properties.branch, { type: "null" }] }
           if (properties?.extra) properties.extra = { anyOf: [properties.extra, { type: "null" }] }
         }
+        if (path === "/experimental/workspace/warp" && method === "post") {
+          const ref = operation.requestBody.content?.["application/json"]?.schema?.$ref?.replace(
+            "#/components/schemas/",
+            "",
+          )
+          const properties = ref
+            ? spec.components?.schemas?.[ref]?.properties
+            : operation.requestBody.content?.["application/json"]?.schema?.properties
+          if (properties?.id) properties.id = { anyOf: [properties.id, { type: "null" }] }
+        }
       }
       for (const response of Object.values(operation.responses ?? {})) {
         for (const content of Object.values(response.content ?? {})) {
diff --git a/packages/opencode/src/server/routes/instance/index.ts b/packages/opencode/src/server/routes/instance/index.ts
index 8d88c676b0..4c81fd3ba1 100644
--- a/packages/opencode/src/server/routes/instance/index.ts
+++ b/packages/opencode/src/server/routes/instance/index.ts
@@ -27,7 +27,7 @@ import { ProviderRoutes } from "./provider"
 import { EventRoutes } from "./event"
 import { SyncRoutes } from "./sync"
 import { InstanceMiddleware } from "./middleware"
-import { jsonRequest } from "./trace"
+import { jsonRequest, runRequest } from "./trace"
 import { ExperimentalHttpApiServer } from "./httpapi/server"
 import { EventPaths } from "./httpapi/event"
 import { ExperimentalPaths } from "./httpapi/groups/experimental"
@@ -40,6 +40,7 @@ import { SyncPaths } from "./httpapi/groups/sync"
 import { TuiPaths } from "./httpapi/groups/tui"
 import { WorkspacePaths } from "./httpapi/groups/workspace"
 import type { CorsOptions } from "@/server/cors"
+import { errors } from "@/server/error"
 
 export const InstanceRoutes = (upgrade: UpgradeWebSocket, opts?: CorsOptions): Hono => {
   const app = new Hono()
@@ -86,7 +87,10 @@ export const InstanceRoutes = (upgrade: UpgradeWebSocket, opts?: CorsOptions): H
     app.get(InstancePaths.path, (c) => handler(c.req.raw, context))
     app.post(InstancePaths.dispose, (c) => handler(c.req.raw, context))
     app.get(InstancePaths.vcs, (c) => handler(c.req.raw, context))
+    app.get(InstancePaths.vcsStatus, (c) => handler(c.req.raw, context))
     app.get(InstancePaths.vcsDiff, (c) => handler(c.req.raw, context))
+    app.get(InstancePaths.vcsDiffRaw, (c) => handler(c.req.raw, context))
+    app.post(InstancePaths.vcsApply, (c) => handler(c.req.raw, context))
     app.get(InstancePaths.command, (c) => handler(c.req.raw, context))
     app.get(InstancePaths.agent, (c) => handler(c.req.raw, context))
     app.get(InstancePaths.skill, (c) => handler(c.req.raw, context))
@@ -288,6 +292,98 @@ export const InstanceRoutes = (upgrade: UpgradeWebSocket, opts?: CorsOptions): H
           return yield* vcs.diff(c.req.valid("query").mode)
         }),
     )
+    .get(
+      "/vcs/status",
+      describeRoute({
+        summary: "Get VCS status",
+        description: "Retrieve changed files in the current working tree without patches.",
+        operationId: "vcs.status",
+        responses: {
+          200: {
+            description: "VCS status",
+            content: {
+              "application/json": {
+                schema: resolver(Vcs.FileStatus.zod.array()),
+              },
+            },
+          },
+        },
+      }),
+      async (c) =>
+        jsonRequest("InstanceRoutes.vcs.status", c, function* () {
+          const vcs = yield* Vcs.Service
+          return yield* vcs.status()
+        }),
+    )
+    .get(
+      "/vcs/diff/raw",
+      describeRoute({
+        summary: "Get raw VCS diff",
+        description: "Retrieve a raw patch for current uncommitted changes.",
+        operationId: "vcs.diff.raw",
+        responses: {
+          200: {
+            description: "Raw VCS diff",
+            content: {
+              "text/x-diff": {
+                schema: resolver(z.string()),
+              },
+            },
+          },
+        },
+      }),
+      async (c) => {
+        const patch = await runRequest(
+          "InstanceRoutes.vcs.diffRaw",
+          c,
+          Vcs.Service.use((vcs) => vcs.diffRaw()),
+        )
+        return c.text(patch, 200, { "content-type": "text/x-diff; charset=utf-8" })
+      },
+    )
+    .post(
+      "/vcs/apply",
+      describeRoute({
+        summary: "Apply VCS patch",
+        description: "Apply a raw patch to the current working tree.",
+        operationId: "vcs.apply",
+        responses: {
+          200: {
+            description: "VCS patch applied",
+            content: {
+              "application/json": {
+                schema: resolver(Vcs.ApplyResult.zod),
+              },
+            },
+          },
+          ...errors(400),
+        },
+      }),
+      validator("json", Vcs.ApplyInput.zodObject),
+      async (c) => {
+        const result = await runRequest(
+          "InstanceRoutes.vcs.apply",
+          c,
+          Vcs.Service.use((vcs) => vcs.apply(c.req.valid("json") as Vcs.ApplyInput)).pipe(
+            Effect.match({
+              onFailure: (error) => ({ ok: false as const, error }),
+              onSuccess: (value) => ({ ok: true as const, value }),
+            }),
+          ),
+        )
+        if (result.ok) return c.json(result.value)
+        return c.json(
+          {
+            name: "VcsApplyError",
+            data: {
+              message: result.error.message,
+              reason: result.error.reason,
+            },
+          },
+          400,
+        )
+      },
+    )
     .get(
       "/command",
       describeRoute({
diff --git a/packages/opencode/src/server/routes/ui.ts b/packages/opencode/src/server/routes/ui.ts
index ce06b2b35e..608525b63a 100644
--- a/packages/opencode/src/server/routes/ui.ts
+++ b/packages/opencode/src/server/routes/ui.ts
@@ -1,10 +1,9 @@
 import fs from "node:fs/promises"
-import { createHash } from "node:crypto"
 import { AppFileSystem } from "@opencode-ai/core/filesystem"
 import { Hono } from "hono"
 import { proxy } from "hono/proxy"
 import { ProxyUtil } from "../proxy-util"
-import { DEFAULT_CSP, UI_UPSTREAM, csp, embeddedUI, themePreloadHash, upstreamURL } from "../shared/ui"
+import { UI_UPSTREAM, csp, cspForHtml, embeddedUI, upstreamURL } from "../shared/ui"
 
 export async function serveUI(request: Request) {
   const embeddedWebUI = await embeddedUI()
@@ -17,8 +16,11 @@ export async function serveUI(request: Request) {
     if (await fs.exists(match)) {
       const mime = AppFileSystem.mimeType(match)
       const headers = new Headers({ "content-type": mime })
-      if (mime.startsWith("text/html")) headers.set("content-security-policy", DEFAULT_CSP)
-      return new Response(new Uint8Array(await fs.readFile(match)), { headers })
+      const body = new Uint8Array(await fs.readFile(match))
+      if (mime.startsWith("text/html")) {
+        headers.set("content-security-policy", cspForHtml(new TextDecoder().decode(body)))
+      }
+      return new Response(body, { headers })
     }
 
     return Response.json({ error: "Not Found" }, { status: 404 })
@@ -28,11 +30,10 @@ export async function serveUI(request: Request) {
     raw: request,
     headers: ProxyUtil.headers(request, { host: UI_UPSTREAM.host }),
   })
-  const match = response.headers.get("content-type")?.includes("text/html")
-    ? themePreloadHash(await response.clone().text())
-    : undefined
-  const hash = match ? createHash("sha256").update(match[2]).digest("base64") : ""
-  response.headers.set("Content-Security-Policy", csp(hash))
+  response.headers.set(
+    "Content-Security-Policy",
+    response.headers.get("content-type")?.includes("text/html") ? cspForHtml(await response.clone().text()) : csp(),
+  )
   return response
 }
 
diff --git a/packages/opencode/src/server/server.ts b/packages/opencode/src/server/server.ts
index b9cc5c8060..d262594755 100644
--- a/packages/opencode/src/server/server.ts
+++ b/packages/opencode/src/server/server.ts
@@ -107,10 +107,10 @@ function createHono(opts: CorsOptions, selection: ServerBackend.Selection = Serv
   const backendAttributes = ServerBackend.attributes(selection)
   const app = new Hono()
     .onError(ErrorMiddleware)
-    .use(AuthMiddleware)
+    .use(CorsMiddleware(opts))
     .use(LoggerMiddleware(backendAttributes))
+    .use(AuthMiddleware)
     .use(CompressionMiddleware)
-    .use(CorsMiddleware(opts))
     .route("/global", GlobalRoutes())
 
   const runtime = adapter.create(app)
diff --git a/packages/opencode/src/server/shared/ui.ts b/packages/opencode/src/server/shared/ui.ts
index 0328663da5..0e27dcf220 100644
--- a/packages/opencode/src/server/shared/ui.ts
+++ b/packages/opencode/src/server/shared/ui.ts
@@ -10,17 +10,21 @@ const embeddedUIPromise = Flag.OPENCODE_DISABLE_EMBEDDED_WEB_UI
   : // @ts-expect-error - generated file at build time
     import("opencode-web-ui.gen.ts").then((module) => module.default as Record<string, string>).catch(() => null)
 
-export const DEFAULT_CSP =
-  "default-src 'self'; script-src 'self' 'wasm-unsafe-eval'; style-src 'self' 'unsafe-inline'; img-src 'self' data: https:; font-src 'self' data:; media-src 'self' data:; connect-src *"
 export const UI_UPSTREAM = new URL("https://app.opencode.ai")
 
 export const csp = (hash = "") =>
-  `default-src 'self'; script-src 'self' 'wasm-unsafe-eval'${hash ? ` 'sha256-${hash}'` : ""}; style-src 'self' 'unsafe-inline'; img-src 'self' data: https:; font-src 'self' data:; media-src 'self' data:; connect-src *`
+  `default-src 'self'; script-src 'self' 'wasm-unsafe-eval'${hash ? ` 'sha256-${hash}'` : ""}; style-src 'self' 'unsafe-inline'; img-src 'self' data: https:; font-src 'self' data:; media-src 'self' data:; connect-src * data:`
+export const DEFAULT_CSP = csp()
 
 export function themePreloadHash(body: string) {
   return body.match(/<script\b(?![^>]*\bsrc\s*=)[^>]*\bid=(['"])oc-theme-preload-script\1[^>]*>([\s\S]*?)<\/script>/i)
 }
 
+export function cspForHtml(body: string) {
+  const match = themePreloadHash(body)
+  return csp(match ? createHash("sha256").update(match[2]).digest("base64") : "")
+}
+
 function requestBody(request: HttpServerRequest.HttpServerRequest) {
   if (request.method === "GET" || request.method === "HEAD") return HttpBody.empty
   const len = request.headers["content-length"]
@@ -53,7 +57,9 @@ function notFound() {
 function embeddedUIResponse(file: string, body: Uint8Array) {
   const mime = AppFileSystem.mimeType(file)
   const headers = new Headers({ "content-type": mime })
-  if (mime.startsWith("text/html")) headers.set("content-security-policy", DEFAULT_CSP)
+  if (mime.startsWith("text/html")) {
+    headers.set("content-security-policy", cspForHtml(new TextDecoder().decode(body)))
+  }
   return HttpServerResponse.raw(body, { headers })
 }
 
@@ -91,8 +97,7 @@ export function serveUIEffect(
 
     if (response.headers["content-type"]?.includes("text/html")) {
       const body = yield* response.text
-      const match = themePreloadHash(body)
-      headers.set("Content-Security-Policy", csp(match ? createHash("sha256").update(match[2]).digest("base64") : ""))
+      headers.set("Content-Security-Policy", cspForHtml(body))
       return HttpServerResponse.text(body, { status: response.status, headers })
     }
 
diff --git a/packages/opencode/src/session/message-v2.ts b/packages/opencode/src/session/message-v2.ts
index 237fb527c0..ed09262d0e 100644
--- a/packages/opencode/src/session/message-v2.ts
+++ b/packages/opencode/src/session/message-v2.ts
@@ -854,13 +854,31 @@ export const toModelMessagesEffect = Effect.fnUntraced(function* (
         role: "assistant",
         parts: [],
       }
+      // Anthropic adaptive thinking can persist assistant turns like:
+      // step-start, reasoning(signature), text(""), step-start,
+      // reasoning(signature). The empty text part is a structural separator,
+      // but it does not carry the signature metadata itself. Dropping it shifts
+      // signed thinking positions after step-start splitting/provider regrouping;
+      // keeping it as "" is filtered by the AI SDK and rejected by Anthropic.
+      // It is unclear whether this shape originates in our stream processing,
+      // a proxy, or a lower-level library, but preserving a non-empty separator
+      // here is the only safe replay point we have.
+      // Use a single space so the separator survives replay without changing
+      // the neighboring signed reasoning blocks. Bedrock-hosted Claude stores
+      // the same signature under the bedrock metadata namespace.
+      const hasSignedReasoning = msg.parts.some((part) => {
+        if (part.type !== "reasoning") return false
+        return part.metadata?.anthropic?.signature != null || part.metadata?.bedrock?.signature != null
+      })
       for (const part of msg.parts) {
-        if (part.type === "text")
+        if (part.type === "text") {
+          const text = part.text === "" && hasSignedReasoning ? " " : part.text
           assistantMessage.parts.push({
             type: "text",
-            text: part.text,
+            text,
             ...(differentModel ? {} : { providerMetadata: part.metadata }),
           })
+        }
         if (part.type === "step-start")
           assistantMessage.parts.push({
             type: "step-start",
diff --git a/packages/opencode/src/session/prompt.ts b/packages/opencode/src/session/prompt.ts
index 8286ecf8e6..fef8c43836 100644
--- a/packages/opencode/src/session/prompt.ts
+++ b/packages/opencode/src/session/prompt.ts
@@ -744,7 +744,7 @@ NOTE: At any point in time through this workflow you should feel free to ask the
           const markReady = ready ? ready.open.pipe(Effect.asVoid) : Effect.void
           const { msg, part, cwd } = yield* Effect.gen(function* () {
             const ctx = yield* InstanceState.context
-            const session = yield* sessions.get(input.sessionID)
+            const session = yield* sessions.get(input.sessionID).pipe(Effect.orDie)
             if (session.revert) {
               yield* revert.cleanup(session)
             }
@@ -1370,7 +1370,7 @@ NOTE: At any point in time through this workflow you should feel free to ask the
 
     const prompt: (input: PromptInput) => Effect.Effect<MessageV2.WithParts> = Effect.fn("SessionPrompt.prompt")(
       function* (input: PromptInput) {
-        const session = yield* sessions.get(input.sessionID)
+        const session = yield* sessions.get(input.sessionID).pipe(Effect.orDie)
         yield* revert.cleanup(session)
         const message = yield* createUserMessage(input)
         yield* sessions.touch(input.sessionID)
@@ -1401,9 +1401,9 @@ NOTE: At any point in time through this workflow you should feel free to ask the
       function* (sessionID: SessionID) {
         const ctx = yield* InstanceState.context
         const slog = elog.with({ sessionID })
-        let structured: unknown | undefined
+        let structured: unknown
         let step = 0
-        const session = yield* sessions.get(sessionID)
+        const session = yield* sessions.get(sessionID).pipe(Effect.orDie)
 
         while (true) {
           yield* status.set(sessionID, { type: "busy" })
diff --git a/packages/opencode/src/session/revert.ts b/packages/opencode/src/session/revert.ts
index 58d69a2040..abf7c3441f 100644
--- a/packages/opencode/src/session/revert.ts
+++ b/packages/opencode/src/session/revert.ts
@@ -44,7 +44,7 @@ export const layer = Layer.effect(
       yield* state.assertNotBusy(input.sessionID)
       const all = yield* sessions.messages({ sessionID: input.sessionID })
       let lastUser: MessageV2.User | undefined
-      const session = yield* sessions.get(input.sessionID)
+      const session = yield* sessions.get(input.sessionID).pipe(Effect.orDie)
 
       let rev: Session.Info["revert"]
       const patches: Snapshot.Patch[] = []
@@ -75,8 +75,8 @@ export const layer = Layer.effect(
       rev.snapshot = session.revert?.snapshot ?? (yield* snap.track())
       if (session.revert?.snapshot) yield* snap.restore(session.revert.snapshot)
       yield* snap.revert(patches)
-      if (rev.snapshot) rev.diff = yield* snap.diff(rev.snapshot as string)
-      const range = all.filter((msg) => msg.info.id >= rev!.messageID)
+      if (rev.snapshot) rev.diff = yield* snap.diff(rev.snapshot)
+      const range = all.filter((msg) => msg.info.id >= rev.messageID)
       const diffs = yield* summary.computeDiff({ messages: range })
       yield* storage.write(["session_diff", input.sessionID], diffs).pipe(Effect.ignore)
       yield* bus.publish(Session.Event.Diff, { sessionID: input.sessionID, diff: diffs })
@@ -89,17 +89,17 @@ export const layer = Layer.effect(
           files: diffs.length,
         },
       })
-      return yield* sessions.get(input.sessionID)
+      return yield* sessions.get(input.sessionID).pipe(Effect.orDie)
     })
 
     const unrevert = Effect.fn("SessionRevert.unrevert")(function* (input: { sessionID: SessionID }) {
       log.info("unreverting", input)
       yield* state.assertNotBusy(input.sessionID)
-      const session = yield* sessions.get(input.sessionID)
+      const session = yield* sessions.get(input.sessionID).pipe(Effect.orDie)
       if (!session.revert) return session
-      if (session.revert.snapshot) yield* snap.restore(session.revert!.snapshot!)
+      if (session.revert.snapshot) yield* snap.restore(session.revert.snapshot)
       yield* sessions.clearRevert(input.sessionID)
-      return yield* sessions.get(input.sessionID)
+      return yield* sessions.get(input.sessionID).pipe(Effect.orDie)
     })
 
     const cleanup = Effect.fn("SessionRevert.cleanup")(function* (session: Session.Info) {
diff --git a/packages/opencode/src/session/session.ts b/packages/opencode/src/session/session.ts
index 2a0c81e8cd..68117ca170 100644
--- a/packages/opencode/src/session/session.ts
+++ b/packages/opencode/src/session/session.ts
@@ -3,7 +3,6 @@ import path from "path"
 import { BusEvent } from "@/bus/bus-event"
 import { Bus } from "@/bus"
 import { Decimal } from "decimal.js"
-import z from "zod"
 import { type ProviderMetadata, type LanguageModelUsage } from "ai"
 import { Flag } from "@opencode-ai/core/flag/flag"
 import { InstallationVersion } from "@opencode-ai/core/installation/version"
@@ -422,6 +421,8 @@ export class BusyError extends Error {
   }
 }
 
+export type NotFound = InstanceType<typeof NotFoundError>
+
 export interface Interface {
   readonly list: (input?: ListInput) => Effect.Effect<Info[]>
   readonly create: (input?: {
@@ -432,9 +433,9 @@ export interface Interface {
     permission?: Permission.Ruleset
     workspaceID?: WorkspaceID
   }) => Effect.Effect<Info>
-  readonly fork: (input: { sessionID: SessionID; messageID?: MessageID }) => Effect.Effect<Info>
+  readonly fork: (input: { sessionID: SessionID; messageID?: MessageID }) => Effect.Effect<Info, NotFound>
   readonly touch: (sessionID: SessionID) => Effect.Effect<void>
-  readonly get: (id: SessionID) => Effect.Effect<Info>
+  readonly get: (id: SessionID) => Effect.Effect<Info, NotFound>
   readonly setTitle: (input: { sessionID: SessionID; title: string }) => Effect.Effect<void>
   readonly setArchived: (input: { sessionID: SessionID; time?: number }) => Effect.Effect<void>
   readonly setPermission: (input: { sessionID: SessionID; permission: Permission.Ruleset }) => Effect.Effect<void>
@@ -448,7 +449,7 @@ export interface Interface {
   readonly diff: (sessionID: SessionID) => Effect.Effect<Snapshot.FileDiff[]>
   readonly messages: (input: { sessionID: SessionID; limit?: number }) => Effect.Effect<MessageV2.WithParts[]>
   readonly children: (parentID: SessionID) => Effect.Effect<Info[]>
-  readonly remove: (sessionID: SessionID) => Effect.Effect<void>
+  readonly remove: (sessionID: SessionID) => Effect.Effect<void, NotFound>
   readonly updateMessage: <T extends MessageV2.Info>(msg: T) => Effect.Effect<T>
   readonly removeMessage: (input: { sessionID: SessionID; messageID: MessageID }) => Effect.Effect<MessageID>
   readonly removePart: (input: { sessionID: SessionID; messageID: MessageID; partID: PartID }) => Effect.Effect<PartID>
@@ -534,13 +535,13 @@ export const layer: Layer.Layer<Service, never, Bus.Service | Storage.Service |
 
     const get = Effect.fn("Session.get")(function* (id: SessionID) {
       const row = yield* db((d) => d.select().from(SessionTable).where(eq(SessionTable.id, id)).get())
-      if (!row) throw new NotFoundError({ message: `Session not found: ${id}` })
+      if (!row) return yield* Effect.fail(new NotFoundError({ message: `Session not found: ${id}` }))
       return fromRow(row)
     })
 
     const list = Effect.fn("Session.list")(function* (input?: ListInput) {
       const ctx = yield* InstanceState.context
-      return Array.from(listByProject({ projectID: ctx.project.id, ...(input ?? {}) }))
+      return Array.from(listByProject({ projectID: ctx.project.id, ...input }))
     })
 
     const children = Effect.fn("Session.children")(function* (parentID: SessionID) {
@@ -555,8 +556,8 @@ export const layer: Layer.Layer<Service, never, Bus.Service | Storage.Service |
     })
 
     const remove: Interface["remove"] = Effect.fnUntraced(function* (sessionID: SessionID) {
+      const session = yield* get(sessionID)
       try {
-        const session = yield* get(sessionID)
         const kids = yield* children(sessionID)
         for (const child of kids) {
           yield* remove(child.id)
diff --git a/packages/opencode/src/util/locale.ts b/packages/opencode/src/util/locale.ts
index 49f60e9311..ec900b4416 100644
--- a/packages/opencode/src/util/locale.ts
+++ b/packages/opencode/src/util/locale.ts
@@ -63,6 +63,11 @@ export function truncate(str: string, len: number): string {
   return str.slice(0, len - 1) + "…"
 }
 
+export function truncateLeft(str: string, len: number): string {
+  if (str.length <= len) return str
+  return "…" + str.slice(-(len - 1))
+}
+
 export function truncateMiddle(str: string, maxLength: number = 35): string {
   if (str.length <= maxLength) return str
 
diff --git a/packages/opencode/test/acp/agent-interface.test.ts b/packages/opencode/test/acp/agent-interface.test.ts
index 9fa67de829..7c4633d7d8 100644
--- a/packages/opencode/test/acp/agent-interface.test.ts
+++ b/packages/opencode/test/acp/agent-interface.test.ts
@@ -34,10 +34,11 @@ describe("acp.agent interface compliance", () => {
     "loadSession",
     "setSessionMode",
     "authenticate",
-    // Unstable - SDK checks these with unstable_ prefix
+    // Capability-gated methods checked by the SDK router
     "listSessions",
+    "resumeSession",
+    "closeSession",
     "unstable_forkSession",
-    "unstable_resumeSession",
     "unstable_setSessionModel",
   ]
 
diff --git a/packages/opencode/test/cli/cmd/tui/dialog-workspace-create.test.ts b/packages/opencode/test/cli/cmd/tui/dialog-workspace-create.test.ts
new file mode 100644
index 0000000000..a32dc61125
--- /dev/null
+++ b/packages/opencode/test/cli/cmd/tui/dialog-workspace-create.test.ts
@@ -0,0 +1,63 @@
+import { describe, expect, test } from "bun:test"
+import { recentConnectedWorkspaces } from "../../../../src/cli/cmd/tui/component/dialog-workspace-create"
+
+describe("recentConnectedWorkspaces", () => {
+  test("returns unique connected workspaces after filtering missing and inactive entries", () => {
+    const workspaces = [
+      { id: "wrk_a", name: "alpha" },
+      { id: "wrk_b", name: "beta" },
+      { id: "wrk_c", name: "gamma" },
+      { id: "wrk_d", name: "delta" },
+      { id: "wrk_e", name: "epsilon" },
+    ]
+    const status = {
+      wrk_a: "connected",
+      wrk_b: "disconnected",
+      wrk_c: "error",
+      wrk_d: "connected",
+      wrk_e: "connected",
+    } as const
+
+    const { recent } = recentConnectedWorkspaces({
+      sessions: [
+        { time: { updated: 900 } },
+        { workspaceID: "wrk_b", time: { updated: 800 } },
+        { workspaceID: "wrk_a", time: { updated: 700 } },
+        { workspaceID: "wrk_a", time: { updated: 600 } },
+        { workspaceID: "wrk_missing", time: { updated: 500 } },
+        { workspaceID: "wrk_c", time: { updated: 400 } },
+        { workspaceID: "wrk_d", time: { updated: 300 } },
+        { workspaceID: "wrk_e", time: { updated: 200 } },
+      ],
+      get: (workspaceID) => workspaces.find((workspace) => workspace.id === workspaceID),
+      status: (workspaceID) => status[workspaceID as keyof typeof status],
+    })
+
+    expect(recent.map((workspace) => workspace.id)).toEqual(["wrk_a", "wrk_d", "wrk_e"])
+  })
+
+  test("omits the active workspace before limiting recent workspaces", () => {
+    const workspaces = [
+      { id: "wrk_a", name: "alpha" },
+      { id: "wrk_b", name: "beta" },
+      { id: "wrk_c", name: "gamma" },
+      { id: "wrk_d", name: "delta" },
+    ]
+
+    const { recent, hasMore } = recentConnectedWorkspaces({
+      sessions: [
+        { workspaceID: "wrk_a", time: { updated: 400 } },
+        { workspaceID: "wrk_b", time: { updated: 300 } },
+        { workspaceID: "wrk_c", time: { updated: 200 } },
+        { workspaceID: "wrk_d", time: { updated: 100 } },
+      ],
+      get: (workspaceID) => workspaces.find((workspace) => workspace.id === workspaceID),
+      status: () => "connected",
+      limit: 3,
+      omitWorkspaceID: "wrk_a",
+    })
+
+    expect(recent.map((workspace) => workspace.id)).toEqual(["wrk_b", "wrk_c", "wrk_d"])
+    expect(hasMore).toBe(false)
+  })
+})
diff --git a/packages/opencode/test/cli/cmd/tui/provider-options.test.ts b/packages/opencode/test/cli/cmd/tui/provider-options.test.ts
new file mode 100644
index 0000000000..39d6398379
--- /dev/null
+++ b/packages/opencode/test/cli/cmd/tui/provider-options.test.ts
@@ -0,0 +1,29 @@
+import { describe, expect, test } from "bun:test"
+import { normalizeCustomProviderID, providerOptions } from "../../../../src/cli/cmd/tui/component/dialog-provider"
+
+describe("providerOptions", () => {
+  test("includes a synthetic Other option for custom providers", () => {
+    expect(providerOptions([{ id: "openai", name: "OpenAI" }]).at(-1)).toMatchObject({
+      title: "Other",
+      description: "Custom provider",
+      category: "Providers",
+    })
+  })
+
+  test("does not use Other as the generic provider category", () => {
+    expect(providerOptions([{ id: "mistral", name: "Mistral" }])[0]?.category).toBe("Providers")
+  })
+
+  test("does not collide with a configured provider named other", () => {
+    const values = providerOptions([{ id: "other", name: "Other Provider" }]).map((option) => option.value)
+    expect(new Set(values).size).toBe(values.length)
+  })
+
+  test("normalizes and validates custom provider ids", () => {
+    expect(normalizeCustomProviderID("  custom-provider  ")).toBe("custom-provider")
+    expect(normalizeCustomProviderID("custom_provider")).toBe("custom_provider")
+    expect(normalizeCustomProviderID("@ai-sdk/custom-provider")).toBe("custom-provider")
+    expect(normalizeCustomProviderID("-custom-provider")).toBeUndefined()
+    expect(normalizeCustomProviderID("Custom Provider")).toBeUndefined()
+  })
+})
diff --git a/packages/opencode/test/cli/tui/editor-context.test.tsx b/packages/opencode/test/cli/tui/editor-context.test.tsx
index 14dead86ac..2c5aa7fa6c 100644
--- a/packages/opencode/test/cli/tui/editor-context.test.tsx
+++ b/packages/opencode/test/cli/tui/editor-context.test.tsx
@@ -59,6 +59,39 @@ function createWebSocketImpl(...sockets: FakeWebSocket[]) {
   } as unknown as typeof WebSocket
 }
 
+function sendSelection(socket: FakeWebSocket, filePath: string, text = "foo") {
+  socket.message(
+    JSON.stringify({
+      jsonrpc: "2.0",
+      method: "selection_changed",
+      params: {
+        text,
+        filePath,
+        selection: {
+          start: { line: 1, character: 1 },
+          end: { line: 1, character: 4 },
+        },
+      },
+    }),
+  )
+}
+
+function expectedSelection(filePath: string, text = "foo") {
+  return {
+    filePath,
+    source: "websocket" as const,
+    ranges: [
+      {
+        text,
+        selection: {
+          start: { line: 1, character: 1 },
+          end: { line: 1, character: 4 },
+        },
+      },
+    ],
+  }
+}
+
 test("useEditorContext reconnect switches editor server by session directory", async () => {
   await using tmp = await tmpdir()
   const startupDirectory = path.join(tmp.path, "startup")
@@ -93,12 +126,18 @@ test("useEditorContext reconnect switches editor server by session directory", a
   await nextTick()
 
   expect(firstSocket.closed).toBeFalse()
+  sendSelection(firstSocket, path.join(startupDirectory, "file.ts"))
+
+  expect(mounted.editor.selection()).toEqual(expectedSelection(path.join(startupDirectory, "file.ts")))
+  expect(mounted.editor.labelState()).toBe("pending")
 
   mounted.editor.reconnect(sessionDirectory)
   await nextTick()
 
   expect(firstSocket.closed).toBeTrue()
   expect(secondSocket.closed).toBeFalse()
+  expect(mounted.editor.selection()).toBeUndefined()
+  expect(mounted.editor.labelState()).toBe("none")
 
   mounted.dispose()
 })
@@ -131,7 +170,7 @@ test("useEditorContext favors configured port over lock files", async () => {
   mounted.dispose()
 })
 
-test("useEditorContext resets selection when reconnecting", async () => {
+test("useEditorContext clears selection when reconnecting", async () => {
   await using tmp = await tmpdir()
   const startupDirectory = path.join(tmp.path, "startup")
   const ideDirectory = path.join(tmp.path, ".claude", "ide")
@@ -169,45 +208,66 @@ test("useEditorContext resets selection when reconnecting", async () => {
       },
     }),
   )
-  socket.message(
-    JSON.stringify({
-      jsonrpc: "2.0",
-      method: "selection_changed",
-      params: {
-        text: "foo",
-        filePath: path.join(startupDirectory, "file.ts"),
-        selection: {
-          start: { line: 1, character: 1 },
-          end: { line: 1, character: 4 },
-        },
-      },
-    }),
-  )
+  sendSelection(socket, path.join(startupDirectory, "file.ts"))
 
   expect(mounted.editor.connected()).toBeTrue()
   expect(mounted.editor.server()).toEqual({
     protocolVersion: "2025-11-25",
     serverInfo: { name: "test", version: "0.0.0" },
   })
-  expect(mounted.editor.selection()).toEqual({
-    filePath: path.join(startupDirectory, "file.ts"),
-    source: "websocket",
-    ranges: [
-      {
-        text: "foo",
-        selection: {
-          start: { line: 1, character: 1 },
-          end: { line: 1, character: 4 },
-        },
-      },
-    ],
-  })
+  expect(mounted.editor.selection()).toEqual(expectedSelection(path.join(startupDirectory, "file.ts")))
+  expect(mounted.editor.labelState()).toBe("pending")
+  mounted.editor.markSelectionSent()
+  expect(mounted.editor.labelState()).toBe("sent")
 
   mounted.editor.reconnect(startupDirectory)
 
   expect(socket.closed).toBeFalse()
   expect(mounted.editor.connected()).toBeTrue()
   expect(mounted.editor.selection()).toBeUndefined()
+  expect(mounted.editor.labelState()).toBe("none")
+
+  mounted.dispose()
+})
+
+test("useEditorContext preserves selection for the next reconnect when requested", async () => {
+  await using tmp = await tmpdir()
+  const startupDirectory = path.join(tmp.path, "startup")
+  const ideDirectory = path.join(tmp.path, ".claude", "ide")
+  await mkdir(startupDirectory, { recursive: true })
+  await mkdir(ideDirectory, { recursive: true })
+  await writeFile(
+    path.join(ideDirectory, "3001.lock"),
+    JSON.stringify({
+      transport: "ws",
+      workspaceFolders: [startupDirectory],
+    }),
+  )
+
+  process.env.CLAUDE_CODE_SSE_PORT = undefined
+  process.env.OPENCODE_EDITOR_SSE_PORT = undefined
+  spyOn(process, "cwd").mockImplementation(() => startupDirectory)
+  spyOn(os, "homedir").mockImplementation(() => tmp.path)
+  const socket = new FakeWebSocket("ws://127.0.0.1:3001")
+
+  const mounted = mountEditorContext(createWebSocketImpl(socket))
+  await nextTick()
+
+  sendSelection(socket, path.join(startupDirectory, "file.ts"))
+  expect(mounted.editor.selection()).toEqual(expectedSelection(path.join(startupDirectory, "file.ts")))
+
+  mounted.editor.markSelectionSent()
+  mounted.editor.preserveSelectionFromNewSession()
+  mounted.editor.reconnect(startupDirectory)
+
+  expect(socket.closed).toBeFalse()
+  expect(mounted.editor.selection()).toEqual(expectedSelection(path.join(startupDirectory, "file.ts")))
+  expect(mounted.editor.labelState()).toBe("sent")
+
+  mounted.editor.reconnect(startupDirectory)
+
+  expect(mounted.editor.selection()).toBeUndefined()
+  expect(mounted.editor.labelState()).toBe("none")
 
   mounted.dispose()
 })
diff --git a/packages/opencode/test/config/config.test.ts b/packages/opencode/test/config/config.test.ts
index 0a522b0850..bbe585237b 100644
--- a/packages/opencode/test/config/config.test.ts
+++ b/packages/opencode/test/config/config.test.ts
@@ -1972,6 +1972,83 @@ test("wellknown URL with trailing slash is normalized", async () => {
   }
 })
 
+test("wellknown remote_config supports templated env vars in headers", async () => {
+  const originalFetch = globalThis.fetch
+  const originalToken = process.env.TEST_TOKEN
+  let wellknownFetchedUrl: string | undefined
+  let remoteFetchedUrl: string | undefined
+  let remoteHeaders: HeadersInit | undefined
+  globalThis.fetch = mock((url: string | URL | Request, init?: RequestInit) => {
+    const urlStr = url instanceof Request ? url.url : url instanceof URL ? url.href : url
+    if (urlStr.includes(".well-known/opencode")) {
+      wellknownFetchedUrl = urlStr
+      return Promise.resolve(
+        new Response(
+          JSON.stringify({
+            remote_config: {
+              url: "https://config.example.com/opencode.json",
+              headers: {
+                Authorization: "Bearer {env:TEST_TOKEN}",
+              },
+            },
+          }),
+          { status: 200 },
+        ),
+      )
+    }
+    if (urlStr.includes("config.example.com")) {
+      remoteFetchedUrl = urlStr
+      remoteHeaders = init?.headers
+      return Promise.resolve(
+        new Response(
+          JSON.stringify({
+            mcp: { confluence: { type: "remote", url: "https://confluence.example.com/mcp", enabled: true } },
+          }),
+          { status: 200 },
+        ),
+      )
+    }
+    return originalFetch(url, init)
+  }) as unknown as typeof fetch
+
+  const fakeAuth = Layer.mock(Auth.Service)({
+    all: () =>
+      Effect.succeed({
+        "https://example.com": new Auth.WellKnown({ type: "wellknown", key: "TEST_TOKEN", token: "test-token" }),
+      }),
+  })
+
+  const layer = Config.layer.pipe(
+    Layer.provide(testFlock),
+    Layer.provide(AppFileSystem.defaultLayer),
+    Layer.provide(Env.defaultLayer),
+    Layer.provide(fakeAuth),
+    Layer.provide(emptyAccount),
+    Layer.provideMerge(infra),
+    Layer.provide(noopNpm),
+  )
+
+  try {
+    await provideTmpdirInstance(
+      () =>
+        Config.Service.use((svc) =>
+          Effect.gen(function* () {
+            const config = yield* svc.get()
+            expect(wellknownFetchedUrl).toBe("https://example.com/.well-known/opencode")
+            expect(remoteFetchedUrl).toBe("https://config.example.com/opencode.json")
+            expect(remoteHeaders).toEqual({ Authorization: "Bearer test-token" })
+            expect(config.mcp?.confluence?.enabled).toBe(true)
+          }),
+        ),
+      { git: true },
+    ).pipe(Effect.scoped, Effect.provide(layer), Effect.runPromise)
+  } finally {
+    globalThis.fetch = originalFetch
+    if (originalToken === undefined) delete process.env.TEST_TOKEN
+    else process.env.TEST_TOKEN = originalToken
+  }
+})
+
 describe("resolvePluginSpec", () => {
   test("keeps package specs unchanged", async () => {
     await using tmp = await tmpdir()
diff --git a/packages/opencode/test/control-plane/workspace.test.ts b/packages/opencode/test/control-plane/workspace.test.ts
index 84f5670064..0eba431e1a 100644
--- a/packages/opencode/test/control-plane/workspace.test.ts
+++ b/packages/opencode/test/control-plane/workspace.test.ts
@@ -1,4 +1,5 @@
 import { afterEach, beforeEach, describe, expect, mock, test } from "bun:test"
+import { $ } from "bun"
 import fs from "node:fs/promises"
 import Http from "node:http"
 import path from "node:path"
@@ -29,12 +30,17 @@ import { WorkspaceTable } from "../../src/control-plane/workspace.sql"
 import type { Target, WorkspaceAdapter, WorkspaceInfo } from "../../src/control-plane/types"
 import * as WorkspaceOld from "../../src/control-plane/workspace"
 import { AppRuntime } from "@/effect/app-runtime"
+import { InstanceStore } from "@/project/instance-store"
+import { InstanceBootstrap } from "@/project/bootstrap"
 
 void Log.init({ print: false })
 
 const testServerLayer = Layer.mergeAll(
   NodeHttpServer.layer(Http.createServer, { host: "127.0.0.1", port: 0 }),
-  WorkspaceOld.defaultLayer,
+  WorkspaceOld.defaultLayer.pipe(
+    Layer.provide(InstanceStore.defaultLayer),
+    Layer.provide(InstanceBootstrap.defaultLayer),
+  ),
   SessionNs.defaultLayer,
 )
 const it = testEffect(testServerLayer)
@@ -107,6 +113,18 @@ async function withInstance<T>(fn: (dir: string) => T | Promise<T>) {
   })
 }
 
+async function initGitRepo(dir: string) {
+  await fs.mkdir(dir, { recursive: true })
+  await $`git init`.cwd(dir).quiet()
+  await $`git config core.fsmonitor false`.cwd(dir).quiet()
+  await $`git config commit.gpgsign false`.cwd(dir).quiet()
+  await $`git config user.email "test@opencode.test"`.cwd(dir).quiet()
+  await $`git config user.name "Test"`.cwd(dir).quiet()
+  await fs.writeFile(path.join(dir, "tracked.txt"), "base\n")
+  await $`git add tracked.txt`.cwd(dir).quiet()
+  await $`git commit -m "base"`.cwd(dir).quiet()
+}
+
 const runWorkspace = <A, E>(effect: Effect.Effect<A, E, WorkspaceOld.Service>) => AppRuntime.runPromise(effect)
 const createWorkspace = (input: WorkspaceOld.CreateInput) =>
   runWorkspace(WorkspaceOld.Service.use((workspace) => workspace.create(input)))
@@ -644,6 +662,33 @@ describe("workspace-old CRUD", () => {
     })
   })
 
+  test("sessionWarp applies source workspace patch to local target workspace", async () => {
+    await withInstance(async (dir) => {
+      const previousType = unique("warp-patch-prev-local")
+      const targetType = unique("warp-patch-target-local")
+      const previousDir = path.join(dir, "warp-patch-prev-local")
+      const targetDir = path.join(dir, "warp-patch-target-local")
+      await initGitRepo(previousDir)
+      await initGitRepo(targetDir)
+      await fs.writeFile(path.join(previousDir, "tracked.txt"), "changed\n")
+      await fs.writeFile(path.join(previousDir, "new.txt"), "new\n")
+
+      const previous = workspaceInfo(Instance.project.id, previousType)
+      const target = workspaceInfo(Instance.project.id, targetType)
+      insertWorkspace(previous)
+      insertWorkspace(target)
+      registerAdapter(Instance.project.id, previousType, localAdapter(previousDir, { createDir: false }).adapter)
+      registerAdapter(Instance.project.id, targetType, localAdapter(targetDir, { createDir: false }).adapter)
+      const session = await AppRuntime.runPromise(SessionNs.Service.use((svc) => svc.create({})))
+      attachSessionToWorkspace(session.id, previous.id)
+
+      await warpWorkspaceSession({ workspaceID: target.id, sessionID: session.id, copyChanges: true })
+
+      expect(await fs.readFile(path.join(targetDir, "tracked.txt"), "utf8")).toBe("changed\n")
+      expect(await fs.readFile(path.join(targetDir, "new.txt"), "utf8")).toBe("new\n")
+    })
+  })
+
   test("sessionWarp detaches a session to the local project and claims project ownership", async () => {
     await withInstance(async (dir) => {
       const previousType = unique("warp-detach-local")
@@ -696,10 +741,12 @@ describe("workspace-old CRUD", () => {
               },
             ])
           }
+          if (call.url.pathname === "/warp-source/vcs/diff/raw") return HttpServerResponse.text("remote patch")
           if (call.url.pathname === "/warp-target/sync/replay")
             return yield* HttpServerResponse.json({ sessionID: "ok" })
           if (call.url.pathname === "/warp-target/sync/steal")
             return yield* HttpServerResponse.json({ sessionID: "ok" })
+          if (call.url.pathname === "/warp-target/vcs/apply") return yield* HttpServerResponse.json({ applied: true })
           return HttpServerResponse.text("unexpected", { status: 500 })
         }),
       )
@@ -722,15 +769,18 @@ describe("workspace-old CRUD", () => {
             historySessionID = session.id
             historyNextSeq = (sessionSequence(session.id) ?? -1) + 1
 
-            yield* workspace.sessionWarp({ workspaceID: target.id, sessionID: session.id })
+            yield* workspace.sessionWarp({ workspaceID: target.id, sessionID: session.id, copyChanges: true })
 
             expect(calls.map((call) => `${call.method} ${call.url.pathname}`)).toEqual([
               "POST /warp-source/sync/history",
+              "GET /warp-source/vcs/diff/raw",
+              "POST /warp-target/vcs/apply",
               "POST /warp-target/sync/replay",
               "POST /warp-target/sync/steal",
             ])
             expect(calls[0].json).toEqual({ [session.id]: historyNextSeq - 1 })
-            expect(calls[1].json).toMatchObject({
+            expect(calls[2].json).toEqual({ patch: "remote patch" })
+            expect(calls[3].json).toMatchObject({
               directory: "remote-target-dir",
               events: [
                 {
@@ -745,7 +795,7 @@ describe("workspace-old CRUD", () => {
                 },
               ],
             })
-            expect(calls[2].json).toEqual({ sessionID: session.id })
+            expect(calls[4].json).toEqual({ sessionID: session.id })
             expect((yield* sessionSvc.get(session.id)).title).toBe("from source history")
             expect(sessionSequenceOwner(session.id)).toBe(target.id)
           }),
@@ -1060,7 +1110,7 @@ describe("workspace-old sync state", () => {
 
               yield* eventuallyEffect(
                 Effect.gen(function* () {
-                  expect((yield* sessionSvc.get(session.id)).title).toBe("from history")
+                  expect((yield* sessionSvc.get(session.id).pipe(Effect.orDie)).title).toBe("from history")
                 }),
               )
               expect(historyBodies).toEqual([{ [session.id]: historyNextSeq - 1 }])
@@ -1208,7 +1258,7 @@ describe("workspace-old sync state", () => {
 
               yield* eventuallyEffect(
                 Effect.gen(function* () {
-                  expect((yield* sessionSvc.get(session.id)).title).toBe("from sse")
+                  expect((yield* sessionSvc.get(session.id).pipe(Effect.orDie)).title).toBe("from sse")
                 }),
               )
               expect(
diff --git a/packages/opencode/test/plugin/workspace-adapter.test.ts b/packages/opencode/test/plugin/workspace-adapter.test.ts
index 249087808d..9199a85a61 100644
--- a/packages/opencode/test/plugin/workspace-adapter.test.ts
+++ b/packages/opencode/test/plugin/workspace-adapter.test.ts
@@ -12,8 +12,14 @@ process.env.OPENCODE_DISABLE_DEFAULT_PLUGINS = "1"
 const { Flag } = await import("@opencode-ai/core/flag/flag")
 const { Plugin } = await import("../../src/plugin/index")
 const { Workspace } = await import("../../src/control-plane/workspace")
+const { InstanceBootstrap } = await import("../../src/project/bootstrap")
 const { Instance } = await import("../../src/project/instance")
-const it = testEffect(Layer.mergeAll(Plugin.defaultLayer, Workspace.defaultLayer, CrossSpawnSpawner.defaultLayer))
+const { InstanceStore } = await import("../../src/project/instance-store")
+const workspaceLayer = Workspace.defaultLayer.pipe(
+  Layer.provide(InstanceStore.defaultLayer),
+  Layer.provide(InstanceBootstrap.defaultLayer),
+)
+const it = testEffect(Layer.mergeAll(Plugin.defaultLayer, workspaceLayer, CrossSpawnSpawner.defaultLayer))
 
 const experimental = Flag.OPENCODE_EXPERIMENTAL_WORKSPACES
 
diff --git a/packages/opencode/test/provider/cf-ai-gateway-e2e.test.ts b/packages/opencode/test/provider/cf-ai-gateway-e2e.test.ts
new file mode 100644
index 0000000000..0c692c50c8
--- /dev/null
+++ b/packages/opencode/test/provider/cf-ai-gateway-e2e.test.ts
@@ -0,0 +1,131 @@
+// End-to-end regression test for opencode#24432.
+//
+// Routes through the actual ai-gateway-provider + @ai-sdk/openai-compatible
+// chain that provider.ts:811 builds at runtime, with only the network boundary
+// stubbed. Asserts that `reasoning_effort` (and other provider options the
+// transform emits) actually land in the body Cloudflare AI Gateway forwards
+// upstream, which is the only place the bug was observable.
+
+import { afterEach, beforeEach, describe, expect, test } from "bun:test"
+import type { JSONValue } from "ai"
+import { generateText } from "ai"
+import { createAiGateway } from "ai-gateway-provider"
+import { createUnified } from "ai-gateway-provider/providers/unified"
+import { ProviderTransform } from "@/provider/transform"
+import type * as Provider from "@/provider/provider"
+import { ModelID, ProviderID } from "@/provider/schema"
+
+type Captured = { url: string; outerBody: unknown }
+type ProviderOptions = Record<string, Record<string, JSONValue>>
+
+const realFetch = globalThis.fetch
+let captured: Captured | null = null
+
+function isRecord(value: unknown): value is Record<string, unknown> {
+  return typeof value === "object" && value !== null && !Array.isArray(value)
+}
+
+beforeEach(() => {
+  captured = null
+  const handle = async (input: Parameters<typeof fetch>[0], init?: Parameters<typeof fetch>[1]): Promise<Response> => {
+    const url = typeof input === "string" ? input : input instanceof URL ? input.toString() : input.url
+    if (url.startsWith("https://gateway.ai.cloudflare.com/")) {
+      const bodyText = typeof init?.body === "string" ? init.body : ""
+      captured = { url, outerBody: bodyText ? JSON.parse(bodyText) : null }
+      return new Response(
+        JSON.stringify({
+          id: "chatcmpl-test",
+          object: "chat.completion",
+          created: 0,
+          model: "openai/gpt-5.4",
+          choices: [{ index: 0, message: { role: "assistant", content: "ok" }, finish_reason: "stop" }],
+          usage: { prompt_tokens: 1, completion_tokens: 1, total_tokens: 2 },
+        }),
+        { status: 200, headers: { "Content-Type": "application/json" } },
+      )
+    }
+    return realFetch(input, init)
+  }
+  // `typeof fetch` includes Bun's `preconnect` method; preserve it from realFetch.
+  const stubFetch: typeof fetch = Object.assign(handle, { preconnect: realFetch.preconnect.bind(realFetch) })
+  globalThis.fetch = stubFetch
+})
+
+afterEach(() => {
+  globalThis.fetch = realFetch
+})
+
+const cfModel = (apiId: string, releaseDate = "2026-03-05"): Provider.Model => ({
+  id: ModelID.make(`cloudflare-ai-gateway/${apiId}`),
+  providerID: ProviderID.make("cloudflare-ai-gateway"),
+  name: apiId,
+  api: { id: apiId, url: "https://gateway.ai.cloudflare.com/v1/compat", npm: "ai-gateway-provider" },
+  capabilities: {
+    reasoning: true,
+    temperature: false,
+    attachment: true,
+    toolcall: true,
+    input: { text: true, audio: false, image: true, video: false, pdf: true },
+    output: { text: true, audio: false, image: false, video: false, pdf: false },
+    interleaved: false,
+  },
+  cost: { input: 1, output: 1, cache: { read: 0, write: 0 } },
+  limit: { context: 1_000_000, output: 128_000 },
+  status: "active",
+  options: {},
+  headers: {},
+  release_date: releaseDate,
+})
+
+// ai-gateway-provider sends an array of step descriptors; each entry's `query`
+// is the body forwarded to the upstream provider.
+function extractUpstreamQuery(body: unknown): Record<string, unknown> | undefined {
+  if (!Array.isArray(body) || body.length === 0) return undefined
+  const first = body[0]
+  if (!isRecord(first)) return undefined
+  const query = first.query
+  return isRecord(query) ? query : undefined
+}
+
+async function callThroughGateway(apiId: string, providerOptions: ProviderOptions) {
+  const aigateway = createAiGateway({ accountId: "test", gateway: "test", apiKey: "test" })
+  const unified = createUnified()
+  await generateText({ model: aigateway(unified(apiId)), prompt: "hi", providerOptions })
+  return extractUpstreamQuery(captured?.outerBody)
+}
+
+describe("cf-ai-gateway end-to-end (regression: #24432)", () => {
+  test("ProviderTransform.providerOptions output puts reasoning_effort on the wire", async () => {
+    // The full chain the runtime exercises:
+    //   transform.providerOptions() -> openaiCompatible key
+    //   -> @ai-sdk/openai-compatible reads it as compatibleOptions
+    //   -> emits body.reasoning_effort
+    //   -> ai-gateway-provider wraps the body and forwards to gateway.ai.cloudflare.com
+    const opts = ProviderTransform.providerOptions(cfModel("openai/gpt-5.4"), { reasoningEffort: "xhigh" })
+    expect(opts).toEqual({ openaiCompatible: { reasoningEffort: "xhigh" } })
+
+    const upstream = await callThroughGateway("openai/gpt-5.4", opts)
+    expect(upstream?.reasoning_effort).toBe("xhigh")
+  })
+
+  test("variants() output for openai/gpt-5.4 lands xhigh on the wire", async () => {
+    // The other half of the bug: workflow `variant: xhigh` flows through variants()
+    // and must reach the wire. variants() returns the providerOptions payload
+    // unwrapped; providerOptions() wraps it under the SDK key.
+    const variants = ProviderTransform.variants(cfModel("openai/gpt-5.4"))
+    expect(variants.xhigh).toEqual({ reasoningEffort: "xhigh" })
+
+    const opts = ProviderTransform.providerOptions(cfModel("openai/gpt-5.4"), variants.xhigh)
+    const upstream = await callThroughGateway("openai/gpt-5.4", opts)
+    expect(upstream?.reasoning_effort).toBe("xhigh")
+  })
+
+  test("legacy buggy key 'cloudflare-ai-gateway' does NOT reach the wire (proves the bug)", async () => {
+    // Sanity: confirms the bug class. If a future change accidentally restores
+    // providerID-keyed providerOptions, this test fails before users notice.
+    const upstream = await callThroughGateway("openai/gpt-5.4", {
+      "cloudflare-ai-gateway": { reasoningEffort: "high" },
+    })
+    expect(upstream?.reasoning_effort).toBeUndefined()
+  })
+})
diff --git a/packages/opencode/test/provider/transform.test.ts b/packages/opencode/test/provider/transform.test.ts
index 9b66eaa77c..c7a321d571 100644
--- a/packages/opencode/test/provider/transform.test.ts
+++ b/packages/opencode/test/provider/transform.test.ts
@@ -1123,6 +1123,118 @@ describe("ProviderTransform.message - DeepSeek reasoning content", () => {
   })
 })
 
+describe("ProviderTransform.message - surrogate sanitization", () => {
+  const model = {
+    id: "test/test-model",
+    providerID: "test",
+    api: {
+      id: "test-model",
+      url: "https://api.test.com",
+      npm: "@ai-sdk/openai-compatible",
+    },
+    name: "Test Model",
+    capabilities: {
+      temperature: true,
+      reasoning: true,
+      attachment: true,
+      toolcall: true,
+      input: { text: true, audio: false, image: true, video: false, pdf: false },
+      output: { text: true, audio: false, image: false, video: false, pdf: false },
+      interleaved: false,
+    },
+    cost: { input: 0.001, output: 0.002, cache: { read: 0.0001, write: 0.0002 } },
+    limit: { context: 128000, output: 8192 },
+    status: "active",
+    options: {},
+    headers: {},
+  } as any
+
+  test("replaces lone surrogates in model-visible text", () => {
+    const lone = "\uD83D"
+    const valid = "🚀"
+    const sanitized = "�"
+    const text = (label: string) => `${label} ${lone} and ${valid}`
+    const expected = (label: string) => `${label} ${sanitized} and ${valid}`
+    const msgs = [
+      { role: "system", content: text("system") },
+      { role: "user", content: text("user string") },
+      {
+        role: "user",
+        content: [
+          { type: "text", text: text("user text") },
+          { type: "image", image: "data:image/png;base64,abcd" },
+        ],
+      },
+      { role: "assistant", content: text("assistant string") },
+      {
+        role: "assistant",
+        content: [
+          { type: "text", text: text("assistant text") },
+          { type: "reasoning", text: text("assistant reasoning") },
+          { type: "tool-call", toolCallId: "call-1", toolName: "Read", input: { filePath: ".opencode/tool/emoji.ts" } },
+          {
+            type: "tool-result",
+            toolCallId: "call-2",
+            toolName: "Read",
+            output: { type: "text", value: text("assistant tool text") },
+          },
+          {
+            type: "tool-result",
+            toolCallId: "call-3",
+            toolName: "Read",
+            output: { type: "error-text", value: text("assistant tool error") },
+          },
+          {
+            type: "tool-result",
+            toolCallId: "call-4",
+            toolName: "Read",
+            output: { type: "content", value: [{ type: "text", text: text("assistant tool content") }] },
+          },
+        ],
+      },
+      {
+        role: "tool",
+        content: [
+          {
+            type: "tool-result",
+            toolCallId: "call-5",
+            toolName: "Read",
+            output: { type: "text", value: text("tool text") },
+          },
+          {
+            type: "tool-result",
+            toolCallId: "call-6",
+            toolName: "Read",
+            output: { type: "error-text", value: text("tool error") },
+          },
+          {
+            type: "tool-result",
+            toolCallId: "call-7",
+            toolName: "Read",
+            output: { type: "content", value: [{ type: "text", text: text("tool content") }] },
+          },
+        ],
+      },
+    ] as any[]
+
+    const result = ProviderTransform.message(msgs, model, {}) as any[]
+
+    expect(result[0].content).toBe(expected("system"))
+    expect(result[1].content).toBe(expected("user string"))
+    expect(result[2].content[0].text).toBe(expected("user text"))
+    expect(result[3].content).toBe(expected("assistant string"))
+    expect(result[4].content[0].text).toBe(expected("assistant text"))
+    expect(result[4].content[1].text).toBe(expected("assistant reasoning"))
+    expect(result[4].content[3].output.value).toBe(expected("assistant tool text"))
+    expect(result[4].content[4].output.value).toBe(expected("assistant tool error"))
+    expect(result[4].content[5].output.value[0].text).toBe(expected("assistant tool content"))
+    expect(result[5].content[0].output.value).toBe(expected("tool text"))
+    expect(result[5].content[1].output.value).toBe(expected("tool error"))
+    expect(result[5].content[2].output.value[0].text).toBe(expected("tool content"))
+    expect(result[2].content[1]).toEqual({ type: "image", image: "data:image/png;base64,abcd" })
+  })
+})
+
 describe("ProviderTransform.message - empty image handling", () => {
   const mockModel = {
     id: "anthropic/claude-3-5-sonnet",
@@ -1993,7 +2105,7 @@ describe("ProviderTransform.message - bedrock caching with non-bedrock providerI
     const msgs = [
       {
         role: "system",
-        content: [{ type: "text", text: "You are a helpful assistant" }],
+        content: "You are a helpful assistant",
       },
       {
         role: "user",
@@ -2007,7 +2119,7 @@ describe("ProviderTransform.message - bedrock caching with non-bedrock providerI
     expect(result[0].providerOptions?.bedrock).toEqual({
       cachePoint: { type: "default" },
     })
-    expect(result[0].content[0].providerOptions?.bedrock).toBeUndefined()
+    expect(result[0].content).toBe("You are a helpful assistant")
   })
 })
 
@@ -2044,7 +2156,7 @@ describe("ProviderTransform.message - cache control on gateway", () => {
     const msgs = [
       {
         role: "system",
-        content: [{ type: "text", text: "You are a helpful assistant" }],
+        content: "You are a helpful assistant",
       },
       {
         role: "user",
@@ -2054,7 +2166,7 @@ describe("ProviderTransform.message - cache control on gateway", () => {
 
     const result = ProviderTransform.message(msgs, model, {}) as any[]
 
-    expect(result[0].content[0].providerOptions).toBeUndefined()
+    expect(result[0].content).toBe("You are a helpful assistant")
     expect(result[0].providerOptions).toBeUndefined()
   })
 
@@ -2883,6 +2995,36 @@ describe("ProviderTransform.variants", () => {
       const result = ProviderTransform.variants(model)
       expect(Object.keys(result)).toEqual(["none", "minimal", "low", "medium", "high", "xhigh"])
     })
+
+    test("dotted gpt-5.x ids include 'minimal' (regression: matcher used to miss gpt-5.4)", () => {
+      const model = createMockModel({
+        id: "gpt-5.4",
+        providerID: "openai",
+        api: {
+          id: "gpt-5.4",
+          url: "https://api.openai.com",
+          npm: "@ai-sdk/openai",
+        },
+        release_date: "2026-03-05",
+      })
+      const result = ProviderTransform.variants(model)
+      expect(Object.keys(result)).toEqual(["none", "minimal", "low", "medium", "high", "xhigh"])
+    })
+
+    test("gpt-50 (lookalike) does not get gpt-5 family treatment", () => {
+      const model = createMockModel({
+        id: "gpt-50",
+        providerID: "openai",
+        api: {
+          id: "gpt-50",
+          url: "https://api.openai.com",
+          npm: "@ai-sdk/openai",
+        },
+        release_date: "2024-01-01",
+      })
+      const result = ProviderTransform.variants(model)
+      expect(Object.keys(result)).toEqual(["low", "medium", "high"])
+    })
   })
 
   describe("@ai-sdk/anthropic", () => {
@@ -3330,4 +3472,83 @@ describe("ProviderTransform.variants", () => {
       expect(result).toEqual({})
     })
   })
+
+  describe("ai-gateway-provider (cloudflare-ai-gateway)", () => {
+    const cfModel = (apiId: string, releaseDate = "2024-01-01") =>
+      createMockModel({
+        id: `cloudflare-ai-gateway/${apiId}`,
+        providerID: "cloudflare-ai-gateway",
+        api: {
+          id: apiId,
+          url: "https://gateway.ai.cloudflare.com/v1/compat",
+          npm: "ai-gateway-provider",
+        },
+        release_date: releaseDate,
+      })
+
+    test("openai gpt-5.4 includes xhigh effort (regression: variant=xhigh used to be silently ignored)", () => {
+      const result = ProviderTransform.variants(cfModel("openai/gpt-5.4", "2026-03-05"))
+      expect(result.xhigh).toEqual({ reasoningEffort: "xhigh" })
+      expect(result.high).toEqual({ reasoningEffort: "high" })
+      expect(Object.keys(result)).toContain("minimal")
+    })
+
+    test("openai gpt-5.2-codex includes xhigh", () => {
+      const result = ProviderTransform.variants(cfModel("openai/gpt-5.2-codex", "2025-12-11"))
+      expect(result.xhigh).toEqual({ reasoningEffort: "xhigh" })
+      expect(Object.keys(result)).toEqual(["low", "medium", "high", "xhigh"])
+    })
+
+    test("openai gpt-4o (no reasoning) returns empty", () => {
+      const model = cfModel("openai/gpt-4o")
+      model.capabilities.reasoning = false
+      const result = ProviderTransform.variants(model)
+      expect(result).toEqual({})
+    })
+
+    test("non-openai upstream falls back to widely-supported OAI efforts", () => {
+      const result = ProviderTransform.variants(cfModel("anthropic/claude-sonnet-4-6"))
+      expect(result).toEqual({
+        low: { reasoningEffort: "low" },
+        medium: { reasoningEffort: "medium" },
+        high: { reasoningEffort: "high" },
+      })
+    })
+  })
+})
+
+describe("ProviderTransform.providerOptions - ai-gateway-provider", () => {
+  const createModel = (overrides: Partial<any> = {}) =>
+    ({
+      id: "cloudflare-ai-gateway/openai/gpt-5.4",
+      providerID: "cloudflare-ai-gateway",
+      api: {
+        id: "openai/gpt-5.4",
+        url: "https://gateway.ai.cloudflare.com/v1/compat",
+        npm: "ai-gateway-provider",
+      },
+      capabilities: {
+        temperature: false,
+        reasoning: true,
+        attachment: true,
+        toolcall: true,
+        input: { text: true, audio: false, image: true, video: false, pdf: true },
+        output: { text: true, audio: false, image: false, video: false, pdf: false },
+        interleaved: false,
+      },
+      cost: { input: 1, output: 1, cache: { read: 0, write: 0 } },
+      limit: { context: 1_000_000, output: 128_000 },
+      status: "active",
+      options: {},
+      headers: {},
+      release_date: "2026-03-05",
+      ...overrides,
+    }) as any
+
+  test("routes options under openaiCompatible (the key @ai-sdk/openai-compatible reads)", () => {
+    // Regression: previously fell back to providerID="cloudflare-ai-gateway",
+    // which @ai-sdk/openai-compatible never reads, silently dropping reasoningEffort.
+    const result = ProviderTransform.providerOptions(createModel(), { reasoningEffort: "high" })
+    expect(result).toEqual({ openaiCompatible: { reasoningEffort: "high" } })
+  })
 })
diff --git a/packages/opencode/test/server/httpapi-cors.test.ts b/packages/opencode/test/server/httpapi-cors.test.ts
index 72265ad9bd..8d7e95dfbf 100644
--- a/packages/opencode/test/server/httpapi-cors.test.ts
+++ b/packages/opencode/test/server/httpapi-cors.test.ts
@@ -63,6 +63,19 @@ describe("HttpApi CORS", () => {
     }),
   )
 
+  it.live("adds CORS headers to legacy unauthorized responses", () =>
+    Effect.gen(function* () {
+      const response = yield* Effect.promise(async () =>
+        Server.Legacy().app.request("/global/config", {
+          headers: { origin: "https://app.opencode.ai" },
+        }),
+      )
+
+      expect(response.status).toBe(401)
+      expect(response.headers.get("access-control-allow-origin")).toBe("https://app.opencode.ai")
+    }),
+  )
+
   it.live("uses custom CORS origins passed to the server", () =>
     Effect.gen(function* () {
       const listener = yield* Effect.acquireRelease(
diff --git a/packages/opencode/test/server/httpapi-instance-context.test.ts b/packages/opencode/test/server/httpapi-instance-context.test.ts
index 410dbe7426..5e00d77708 100644
--- a/packages/opencode/test/server/httpapi-instance-context.test.ts
+++ b/packages/opencode/test/server/httpapi-instance-context.test.ts
@@ -10,8 +10,10 @@ import { registerAdapter } from "../../src/control-plane/adapters"
 import type { WorkspaceAdapter } from "../../src/control-plane/types"
 import { Workspace } from "../../src/control-plane/workspace"
 import { InstanceRef, WorkspaceRef } from "../../src/effect/instance-ref"
+import { InstanceBootstrap } from "../../src/project/bootstrap"
 import { Instance } from "../../src/project/instance"
 import { InstanceLayer } from "../../src/project/instance-layer"
+import { InstanceStore } from "../../src/project/instance-store"
 import { Project } from "../../src/project/project"
 import { disposeMiddleware, markInstanceForDisposal } from "../../src/server/routes/instance/httpapi/lifecycle"
 import { instanceRouterMiddleware } from "../../src/server/routes/instance/httpapi/middleware/instance-context"
@@ -36,6 +38,11 @@ const testStateLayer = Layer.effectDiscard(
   }),
 )
 
+const workspaceLayer = Workspace.defaultLayer.pipe(
+  Layer.provide(InstanceStore.defaultLayer),
+  Layer.provide(InstanceBootstrap.defaultLayer),
+)
+
 const it = testEffect(
   Layer.mergeAll(
     testStateLayer,
@@ -43,7 +50,7 @@ const it = testEffect(
     NodeServices.layer,
     InstanceLayer.layer,
     Project.defaultLayer,
-    Workspace.defaultLayer,
+    workspaceLayer,
   ),
 )
 
diff --git a/packages/opencode/test/server/httpapi-parity.test.ts b/packages/opencode/test/server/httpapi-parity.test.ts
index 6922d8c43f..9d7eff4964 100644
--- a/packages/opencode/test/server/httpapi-parity.test.ts
+++ b/packages/opencode/test/server/httpapi-parity.test.ts
@@ -105,23 +105,22 @@ describe("404 mapping for missing session", () => {
 })
 
 // ──────────────────────────────────────────────────────────────────────────────
-// Reproducer 3: 404 response body shape should match Hono's NamedError
-// envelope `{ name, data: { message } }`. HttpApi returns the typed-error
-// shape `{ _tag }` instead. SDK consumers reading `error.data.message`
-// see undefined.
-//
-// FIXME: unskip when error JSON shape policy is decided + applied (separate PR).
+// Reproducer 3: 404 response body shape should match Hono's public NamedError
+// envelope `{ name, data: { message } }`. SDK consumers read
+// `error.data.message`, so returning an Effect built-in `{ _tag }` body is a
+// compatibility break.
 // ──────────────────────────────────────────────────────────────────────────────
 describe("Error JSON shape parity", () => {
-  test.todo("HttpApi 404 body matches NamedError shape", async () => {
+  test("HttpApi 404 body matches Hono shape", async () => {
     await using tmp = await tmpdir({ config: { formatter: false, lsp: false } })
+    const headers = { "x-opencode-directory": tmp.path }
 
-    const response = await app(true).request("/session/ses_does_not_exist", {
-      headers: { "x-opencode-directory": tmp.path },
-    })
+    const hono = await app(false).request("/session/ses_does_not_exist", { headers })
+    const httpapi = await app(true).request("/session/ses_does_not_exist", { headers })
 
-    expect(response.status).toBe(404)
-    const body = (await response.json()) as { name?: string; data?: { message?: string } }
+    expect(httpapi.status).toBe(hono.status)
+    const body = (await httpapi.json()) as { name?: string; data?: { message?: string } }
+    expect(body).toEqual(await hono.json())
     expect(body.name).toBe("NotFoundError")
     expect(typeof body.data?.message).toBe("string")
   })
diff --git a/packages/opencode/test/server/httpapi-pty.test.ts b/packages/opencode/test/server/httpapi-pty.test.ts
index 2b6284a310..5e63eae61c 100644
--- a/packages/opencode/test/server/httpapi-pty.test.ts
+++ b/packages/opencode/test/server/httpapi-pty.test.ts
@@ -50,9 +50,9 @@ const effectIt = testEffect(
   ),
 )
 
-function app() {
-  Flag.OPENCODE_EXPERIMENTAL_HTTPAPI = true
-  return Server.Default().app
+function app(experimental = true) {
+  Flag.OPENCODE_EXPERIMENTAL_HTTPAPI = experimental
+  return experimental ? Server.Default().app : Server.Legacy().app
 }
 
 function serverUrl() {
@@ -121,6 +121,18 @@ describe("pty HttpApi bridge", () => {
     expect(missing.status).toBe(404)
   })
 
+  test("matches Hono missing PTY error body", async () => {
+    await using tmp = await tmpdir({ git: true, config: { formatter: false, lsp: false } })
+    const headers = { "x-opencode-directory": tmp.path }
+    const path = PtyPaths.get.replace(":ptyID", PtyID.ascending())
+
+    const hono = await app(false).request(path, { headers })
+    const httpapi = await app().request(path, { headers })
+
+    expect(httpapi.status).toBe(hono.status)
+    expect(await httpapi.json()).toEqual(await hono.json())
+  })
+
   test("returns 404 for missing PTY websocket before upgrade", async () => {
     await using tmp = await tmpdir({ git: true, config: { formatter: false, lsp: false } })
     const response = await app().request(PtyPaths.connect.replace(":ptyID", PtyID.ascending()), {
diff --git a/packages/opencode/test/server/httpapi-sdk.test.ts b/packages/opencode/test/server/httpapi-sdk.test.ts
index ce774ccfd0..6d2df45078 100644
--- a/packages/opencode/test/server/httpapi-sdk.test.ts
+++ b/packages/opencode/test/server/httpapi-sdk.test.ts
@@ -4,6 +4,7 @@ import type * as Scope from "effect/Scope"
 import { HttpRouter } from "effect/unstable/http"
 import { Flag } from "@opencode-ai/core/flag/flag"
 import { createOpencodeClient } from "@opencode-ai/sdk/v2"
+import { validateSession } from "../../src/cli/cmd/tui/validate-session"
 import { Instance } from "../../src/project/instance"
 import { WithInstance } from "../../src/project/with-instance"
 import { ExperimentalHttpApiServer } from "../../src/server/routes/instance/httpapi/server"
@@ -13,6 +14,7 @@ import { MessageV2 } from "../../src/session/message-v2"
 import { ModelID, ProviderID } from "../../src/provider/schema"
 import type { Config } from "@/config/config"
 import { Session as SessionNs } from "@/session/session"
+import { errorMessage } from "../../src/util/error"
 import { TestLLMServer } from "../lib/llm-server"
 import path from "path"
 import { resetDatabase } from "../fixture/db"
@@ -64,20 +66,23 @@ function client(
   directory?: string,
   input?: { password?: string; username?: string; headers?: Record<string, string> },
 ) {
-  const serverApp = app(backend, input)
-  const fetch = Object.assign(
-    async (request: RequestInfo | URL, init?: RequestInit) =>
-      await serverApp.fetch(request instanceof Request ? request : new Request(request, init)),
-    { preconnect: globalThis.fetch.preconnect },
-  ) satisfies typeof globalThis.fetch
   return createOpencodeClient({
     baseUrl: "http://localhost",
     directory,
     headers: input?.headers,
-    fetch,
+    fetch: serverFetch(backend, input),
   })
 }
 
+function serverFetch(backend: Backend, input?: { password?: string; username?: string }) {
+  const serverApp = app(backend, input)
+  return Object.assign(
+    async (request: RequestInfo | URL, init?: RequestInit) =>
+      await serverApp.fetch(request instanceof Request ? request : new Request(request, init)),
+    { preconnect: globalThis.fetch.preconnect },
+  ) satisfies typeof globalThis.fetch
+}
+
 function authorization(username: string, password: string) {
   return `Basic ${Buffer.from(`${username}:${password}`).toString("base64")}`
 }
@@ -129,6 +134,16 @@ function capture(request: () => Promise<SdkResult>) {
   )
 }
 
+function captureThrown(request: () => Promise<unknown>) {
+  return call(async () => {
+    try {
+      await request()
+    } catch (error) {
+      return error
+    }
+  })
+}
+
 function expectStatus(request: () => Promise<{ response: Response }>, status: number) {
   return call(request).pipe(
     Effect.tap((result) => Effect.sync(() => expect(result.response.status).toBe(status))),
@@ -338,6 +353,46 @@ describe("HttpApi SDK", () => {
     ),
   )
 
+  parity("matches generated SDK missing session errors across backends", (backend) =>
+    withStandardProject(backend, ({ sdk }) =>
+      Effect.gen(function* () {
+        const sessionID = "ses_missing"
+        const expected = {
+          name: "NotFoundError",
+          data: { message: `Session not found: ${sessionID}` },
+        }
+        const missing = yield* capture(() => sdk.session.get({ sessionID }))
+        const thrown = yield* captureThrown(() => sdk.session.get({ sessionID }, { throwOnError: true }))
+
+        expect(missing.error).toEqual(expected)
+        expect(thrown).toEqual(expected)
+        return {
+          status: missing.status,
+          error: missing.error,
+          thrown,
+        }
+      }),
+    ),
+  )
+
+  parity("formats missing session validation errors for -s", (backend) =>
+    withStandardProject(backend, ({ directory }) =>
+      Effect.gen(function* () {
+        const sessionID = "ses_206f84f18ffeZ6hhD7pFYAiW5T"
+        const thrown = yield* captureThrown(() =>
+          validateSession({
+            url: "http://localhost",
+            directory,
+            sessionID,
+            fetch: serverFetch(backend),
+          }),
+        )
+        expect(errorMessage(thrown)).toBe(`Session not found: ${sessionID}`)
+        return errorMessage(thrown)
+      }),
+    ),
+  )
+
   parity("matches generated SDK basic auth behavior across backends", (backend) =>
     withStandardProject(backend, ({ directory }) =>
       Effect.gen(function* () {
diff --git a/packages/opencode/test/server/httpapi-session.test.ts b/packages/opencode/test/server/httpapi-session.test.ts
index 34cecd80d0..c1d82446b9 100644
--- a/packages/opencode/test/server/httpapi-session.test.ts
+++ b/packages/opencode/test/server/httpapi-session.test.ts
@@ -1,20 +1,21 @@
 import { afterEach, describe, expect } from "bun:test"
 import { mkdir } from "node:fs/promises"
 import path from "node:path"
-import { Effect } from "effect"
+import { Effect, Layer } from "effect"
 import { Flag } from "@opencode-ai/core/flag/flag"
 import { registerAdapter } from "../../src/control-plane/adapters"
 import type { WorkspaceAdapter } from "../../src/control-plane/types"
 import { Workspace } from "../../src/control-plane/workspace"
 import { PermissionID } from "../../src/permission/schema"
 import { ModelID, ProviderID } from "../../src/provider/schema"
-import { Instance } from "../../src/project/instance"
 import { WithInstance } from "../../src/project/with-instance"
+import { InstanceBootstrap } from "../../src/project/bootstrap"
+import { InstanceStore } from "../../src/project/instance-store"
 import { Project } from "../../src/project/project"
 import { Server } from "../../src/server/server"
 import { SessionPaths } from "../../src/server/routes/instance/httpapi/groups/session"
 import { Session } from "@/session/session"
-import { MessageID, PartID, type SessionID } from "../../src/session/schema"
+import { MessageID, PartID, SessionID, type SessionID as SessionIDType } from "../../src/session/schema"
 import { MessageV2 } from "../../src/session/message-v2"
 import { Database } from "@/storage/db"
 import { SessionMessageTable, SessionTable } from "@/session/session.sql"
@@ -31,6 +32,10 @@ void Log.init({ print: false })
 
 const original = Flag.OPENCODE_EXPERIMENTAL_HTTPAPI
 const originalWorkspaces = Flag.OPENCODE_EXPERIMENTAL_WORKSPACES
+const workspaceLayer = Workspace.defaultLayer.pipe(
+  Layer.provide(InstanceStore.defaultLayer),
+  Layer.provide(InstanceBootstrap.defaultLayer),
+)
 
 function app(experimental = true) {
   Flag.OPENCODE_EXPERIMENTAL_HTTPAPI = experimental
@@ -55,7 +60,7 @@ function createSession(directory: string, input?: Session.CreateInput) {
   )
 }
 
-function createTextMessage(directory: string, sessionID: SessionID, text: string) {
+function createTextMessage(directory: string, sessionID: SessionIDType, text: string) {
   return Effect.promise(
     async () =>
       await WithInstance.provide({
@@ -107,7 +112,7 @@ const createLocalWorkspace = (input: { projectID: Project.Info["id"]; type: stri
         extra: null,
         projectID: input.projectID,
       }),
-    ).pipe(Effect.provide(Workspace.defaultLayer))
+    ).pipe(Effect.provide(workspaceLayer))
   })
 
 function request(path: string, init?: RequestInit) {
@@ -125,6 +130,10 @@ function json<T>(response: Response) {
   })
 }
 
+function responseJson(response: Response) {
+  return Effect.promise(() => response.json())
+}
+
 function requestJson<T>(path: string, init?: RequestInit) {
   return request(path, init).pipe(Effect.flatMap(json<T>))
 }
@@ -147,6 +156,47 @@ afterEach(async () => {
 })
 
 describe("session HttpApi", () => {
+  it.live(
+    "returns declared not found errors for read routes",
+    withTmp({ git: true, config: { formatter: false, lsp: false } }, (tmp) =>
+      Effect.gen(function* () {
+        const headers = { "x-opencode-directory": tmp.path }
+        const missingSession = SessionID.descending()
+        const missingSessionBody = {
+          name: "NotFoundError",
+          data: { message: `Session not found: ${missingSession}` },
+        }
+
+        const get = yield* request(pathFor(SessionPaths.get, { sessionID: missingSession }), { headers })
+        expect(get.status).toBe(404)
+        expect(yield* responseJson(get)).toEqual(missingSessionBody)
+
+        const messages = yield* request(pathFor(SessionPaths.messages, { sessionID: missingSession }), { headers })
+        expect(messages.status).toBe(404)
+        expect(yield* responseJson(messages)).toEqual(missingSessionBody)
+
+        const remove = yield* request(pathFor(SessionPaths.remove, { sessionID: missingSession }), {
+          headers,
+          method: "DELETE",
+        })
+        expect(remove.status).toBe(404)
+        expect(yield* responseJson(remove)).toEqual(missingSessionBody)
+
+        const session = yield* createSession(tmp.path, { title: "missing message" })
+        const missingMessage = MessageID.ascending()
+        const message = yield* request(
+          pathFor(SessionPaths.message, { sessionID: session.id, messageID: missingMessage }),
+          { headers },
+        )
+        expect(message.status).toBe(404)
+        expect(yield* responseJson(message)).toEqual({
+          name: "NotFoundError",
+          data: { message: `Message not found: ${missingMessage}` },
+        })
+      }),
+    ),
+  )
+
   it.live(
     "serves read routes through Hono bridge",
     withTmp({ git: true, config: { formatter: false, lsp: false } }, (tmp) =>
diff --git a/packages/opencode/test/server/httpapi-tui.test.ts b/packages/opencode/test/server/httpapi-tui.test.ts
index 8d2670c492..91cad362a9 100644
--- a/packages/opencode/test/server/httpapi-tui.test.ts
+++ b/packages/opencode/test/server/httpapi-tui.test.ts
@@ -72,14 +72,27 @@ describe("tui HttpApi bridge", () => {
       properties: { text: "from publish" },
     })
 
+    const missingSessionID = SessionID.descending()
     const missing = await app().request(TuiPaths.selectSession, {
       method: "POST",
       headers: { ...headers, "content-type": "application/json" },
-      body: JSON.stringify({ sessionID: SessionID.descending() }),
+      body: JSON.stringify({ sessionID: missingSessionID }),
     })
     expect(missing.status).toBe(404)
   })
 
+  test("matches Hono missing selected session error body", async () => {
+    await using tmp = await tmpdir({ git: true, config: { formatter: false, lsp: false } })
+    const headers = { "x-opencode-directory": tmp.path, "content-type": "application/json" }
+    const body = JSON.stringify({ sessionID: SessionID.descending() })
+
+    const hono = await app(false).request(TuiPaths.selectSession, { method: "POST", headers, body })
+    const httpapi = await app().request(TuiPaths.selectSession, { method: "POST", headers, body })
+
+    expect(httpapi.status).toBe(hono.status)
+    expect(await httpapi.json()).toEqual(await hono.json())
+  })
+
   test("matches legacy unknown execute command behavior", async () => {
     await using tmp = await tmpdir({ git: true, config: { formatter: false, lsp: false } })
     const headers = { "x-opencode-directory": tmp.path, "content-type": "application/json" }
diff --git a/packages/opencode/test/server/httpapi-ui.test.ts b/packages/opencode/test/server/httpapi-ui.test.ts
index 85162f6a92..440aeaecb5 100644
--- a/packages/opencode/test/server/httpapi-ui.test.ts
+++ b/packages/opencode/test/server/httpapi-ui.test.ts
@@ -1,3 +1,4 @@
+import { createHash } from "node:crypto"
 import { afterEach, describe, expect, test } from "bun:test"
 import { Flag } from "@opencode-ai/core/flag/flag"
 import * as Log from "@opencode-ai/core/util/log"
@@ -260,6 +261,38 @@ describe("HttpApi UI fallback", () => {
     expect(await response.text()).toBe("console.log('embedded')")
   })
 
+  test("allows embedded UI terminal wasm and theme preload CSP", async () => {
+    Flag.OPENCODE_EXPERIMENTAL_HTTPAPI = true
+    const script = 'document.documentElement.dataset.theme = "dark"'
+
+    const response = await Effect.runPromise(
+      Effect.gen(function* () {
+        const fs = yield* AppFileSystem.Service
+        return yield* serveEmbeddedUIEffect(
+          "/",
+          {
+            ...fs,
+            readFile: (path) => {
+              return path === "/$bunfs/root/index.html"
+                ? Effect.succeed(
+                    new TextEncoder().encode(
+                      `<html><head><script id="oc-theme-preload-script">${script}</script></head></html>`,
+                    ),
+                  )
+                : Effect.die(`unexpected embedded UI path: ${path}`)
+            },
+          },
+          { "index.html": "/$bunfs/root/index.html" },
+        )
+      }).pipe(Effect.provide(AppFileSystem.defaultLayer), Effect.map(HttpServerResponse.toWeb)),
+    )
+
+    const csp = response.headers.get("content-security-policy") ?? ""
+    expect(csp).toContain("script-src 'self' 'wasm-unsafe-eval'")
+    expect(csp).toContain(`'sha256-${createHash("sha256").update(script).digest("base64")}'`)
+    expect(csp).toContain("connect-src * data:")
+  })
+
   test("keeps matched API routes ahead of the UI fallback", async () => {
     Flag.OPENCODE_EXPERIMENTAL_HTTPAPI = true
 
diff --git a/packages/opencode/test/server/httpapi-workspace-routing.test.ts b/packages/opencode/test/server/httpapi-workspace-routing.test.ts
index b0b276841d..379b71a91e 100644
--- a/packages/opencode/test/server/httpapi-workspace-routing.test.ts
+++ b/packages/opencode/test/server/httpapi-workspace-routing.test.ts
@@ -20,6 +20,8 @@ import { WorkspaceID } from "../../src/control-plane/schema"
 import type { WorkspaceAdapter } from "../../src/control-plane/types"
 import { Workspace } from "../../src/control-plane/workspace"
 import { WorkspaceTable } from "../../src/control-plane/workspace.sql"
+import { InstanceBootstrap } from "../../src/project/bootstrap"
+import { InstanceStore } from "../../src/project/instance-store"
 import { Project } from "../../src/project/project"
 import { WorkspacePaths } from "../../src/server/routes/instance/httpapi/groups/workspace"
 import {
@@ -45,13 +47,18 @@ const testStateLayer = Layer.effectDiscard(
   }),
 )
 
+const workspaceLayer = Workspace.defaultLayer.pipe(
+  Layer.provide(InstanceStore.defaultLayer),
+  Layer.provide(InstanceBootstrap.defaultLayer),
+)
+
 const it = testEffect(
   Layer.mergeAll(
     testStateLayer,
     NodeHttpServer.layerTest,
     NodeServices.layer,
     Project.defaultLayer,
-    Workspace.defaultLayer,
+    workspaceLayer,
     Socket.layerWebSocketConstructorGlobal,
   ),
 )
diff --git a/packages/opencode/test/server/httpapi-workspace.test.ts b/packages/opencode/test/server/httpapi-workspace.test.ts
index 21bf4120c9..9b38cb44a2 100644
--- a/packages/opencode/test/server/httpapi-workspace.test.ts
+++ b/packages/opencode/test/server/httpapi-workspace.test.ts
@@ -14,6 +14,8 @@ import { Server } from "../../src/server/server"
 import { resetDatabase } from "../fixture/db"
 import { disposeAllInstances, provideInstance, tmpdirScoped } from "../fixture/fixture"
 import { Instance } from "../../src/project/instance"
+import { InstanceBootstrap } from "../../src/project/bootstrap"
+import { InstanceStore } from "../../src/project/instance-store"
 import { Project } from "../../src/project/project"
 import { InstancePaths } from "../../src/server/routes/instance/httpapi/groups/instance"
 import { WorkspaceRef } from "../../src/effect/instance-ref"
@@ -23,9 +25,11 @@ void Log.init({ print: false })
 
 const originalWorkspaces = Flag.OPENCODE_EXPERIMENTAL_WORKSPACES
 const originalHttpApi = Flag.OPENCODE_EXPERIMENTAL_HTTPAPI
-const it = testEffect(
-  Layer.mergeAll(NodeServices.layer, Project.defaultLayer, Session.defaultLayer, Workspace.defaultLayer),
+const workspaceLayer = Workspace.defaultLayer.pipe(
+  Layer.provide(InstanceStore.defaultLayer),
+  Layer.provide(InstanceBootstrap.defaultLayer),
 )
+const it = testEffect(Layer.mergeAll(NodeServices.layer, Project.defaultLayer, Session.defaultLayer, workspaceLayer))
 
 function request(path: string, directory: string, init: RequestInit = {}, httpApi = true) {
   return Effect.promise(() => {
diff --git a/packages/opencode/test/session/message-v2.test.ts b/packages/opencode/test/session/message-v2.test.ts
index a7853be0b8..999b61b48e 100644
--- a/packages/opencode/test/session/message-v2.test.ts
+++ b/packages/opencode/test/session/message-v2.test.ts
@@ -1098,6 +1098,108 @@ describe("session.message-v2.toModelMessage", () => {
       },
     ])
   })
+
+  test("substitutes space for empty text between signed reasoning blocks", async () => {
+    // Reproduces the bug pattern: [reasoning(sig), text(""), reasoning(sig), text(full)]
+    const assistantID = "m-assistant"
+    const input: MessageV2.WithParts[] = [
+      {
+        info: assistantInfo(assistantID, "m-parent"),
+        parts: [
+          { ...basePart(assistantID, "p1"), type: "step-start" },
+          {
+            ...basePart(assistantID, "p2"),
+            type: "reasoning",
+            text: "thinking-one",
+            metadata: { anthropic: { signature: "sig1" } },
+          },
+          { ...basePart(assistantID, "p3"), type: "text", text: "" },
+          { ...basePart(assistantID, "p4"), type: "step-start" },
+          {
+            ...basePart(assistantID, "p5"),
+            type: "reasoning",
+            text: "thinking-two",
+            metadata: { anthropic: { signature: "sig2" } },
+          },
+          { ...basePart(assistantID, "p6"), type: "text", text: "the answer" },
+        ] as MessageV2.Part[],
+      },
+    ]
+
+    const result = await MessageV2.toModelMessages(input, model)
+
+    // step-start splits into two assistant messages; SDK's groupIntoBlocks merges them later
+    expect(result).toHaveLength(2)
+    expect((result[0].content as any[]).find((p) => p.type === "text").text).toBe(" ")
+    expect((result[1].content as any[]).find((p) => p.type === "text").text).toBe("the answer")
+  })
+
+  test("substitutes space for empty text when reasoning signature is under 'bedrock' namespace", async () => {
+    // AWS Bedrock hosts Anthropic Claude but stores signatures under metadata.bedrock
+    const assistantID = "m-assistant-bedrock"
+    const input: MessageV2.WithParts[] = [
+      {
+        info: assistantInfo(assistantID, "m-parent"),
+        parts: [
+          {
+            ...basePart(assistantID, "p1"),
+            type: "reasoning",
+            text: "thinking-bedrock",
+            metadata: { bedrock: { signature: "bedrock-sig" } },
+          },
+          { ...basePart(assistantID, "p2"), type: "text", text: "" },
+          { ...basePart(assistantID, "p3"), type: "text", text: "answer" },
+        ] as MessageV2.Part[],
+      },
+    ]
+
+    const result = await MessageV2.toModelMessages(input, model)
+
+    expect(result).toHaveLength(1)
+    const texts = (result[0].content as any[]).filter((p) => p.type === "text")
+    expect(texts.map((t) => t.text)).toStrictEqual([" ", "answer"])
+  })
+
+  test("leaves empty text alone when reasoning has no Anthropic signature", async () => {
+    // Non-Anthropic providers' reasoning doesn't position-validate, so empty text
+    // should be filtered normally rather than substituted.
+    const assistantID = "m-assistant-unsigned"
+    const input: MessageV2.WithParts[] = [
+      {
+        info: assistantInfo(assistantID, "m-parent"),
+        parts: [
+          { ...basePart(assistantID, "p1"), type: "reasoning", text: "thinking" },
+          { ...basePart(assistantID, "p2"), type: "text", text: "" },
+          { ...basePart(assistantID, "p3"), type: "text", text: "answer" },
+        ] as MessageV2.Part[],
+      },
+    ]
+
+    const result = await MessageV2.toModelMessages(input, model)
+
+    expect(result).toHaveLength(1)
+    const texts = (result[0].content as any[]).filter((p) => p.type === "text")
+    expect(texts.map((t) => t.text)).toStrictEqual(["", "answer"])
+  })
+
+  test("leaves empty text alone in assistant messages without reasoning", async () => {
+    const assistantID = "m-assistant-no-reasoning"
+    const input: MessageV2.WithParts[] = [
+      {
+        info: assistantInfo(assistantID, "m-parent"),
+        parts: [
+          { ...basePart(assistantID, "p1"), type: "text", text: "" },
+          { ...basePart(assistantID, "p2"), type: "text", text: "hello" },
+        ] as MessageV2.Part[],
+      },
+    ]
+
+    const result = await MessageV2.toModelMessages(input, model)
+
+    expect(result).toHaveLength(1)
+    const texts = (result[0].content as any[]).filter((p) => p.type === "text")
+    expect(texts.map((t) => t.text)).toStrictEqual(["", "hello"])
+  })
 })
 
 describe("session.message-v2.fromError", () => {
diff --git a/packages/plugin/package.json b/packages/plugin/package.json
index 9bcf2a6f1f..861208770c 100644
--- a/packages/plugin/package.json
+++ b/packages/plugin/package.json
@@ -1,7 +1,7 @@
 {
   "$schema": "https://json.schemastore.org/package.json",
   "name": "@opencode-ai/plugin",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "type": "module",
   "license": "MIT",
   "scripts": {
diff --git a/packages/sdk/js/package.json b/packages/sdk/js/package.json
index 4ec95155c3..2959cba2dd 100644
--- a/packages/sdk/js/package.json
+++ b/packages/sdk/js/package.json
@@ -1,7 +1,7 @@
 {
   "$schema": "https://json.schemastore.org/package.json",
   "name": "@opencode-ai/sdk",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "type": "module",
   "license": "MIT",
   "scripts": {
diff --git a/packages/sdk/js/src/v2/gen/sdk.gen.ts b/packages/sdk/js/src/v2/gen/sdk.gen.ts
index ffc0970c0e..ebedb1dd6b 100644
--- a/packages/sdk/js/src/v2/gen/sdk.gen.ts
+++ b/packages/sdk/js/src/v2/gen/sdk.gen.ts
@@ -131,6 +131,7 @@ import type {
   SessionDeleteResponses,
   SessionDelivery,
   SessionDiffResponses,
+  SessionForkErrors,
   SessionForkResponses,
   SessionGetErrors,
   SessionGetResponses,
@@ -201,8 +202,12 @@ import type {
   V2SessionMessagesResponses,
   V2SessionPromptResponses,
   V2SessionWaitResponses,
+  VcsApplyErrors,
+  VcsApplyResponses,
+  VcsDiffRawResponses,
   VcsDiffResponses,
   VcsGetResponses,
+  VcsStatusResponses,
   WorktreeCreateErrors,
   WorktreeCreateInput,
   WorktreeCreateResponses,
@@ -1019,8 +1024,9 @@ export class Workspace extends HeyApiClient {
     parameters?: {
       directory?: string
       workspace?: string
-      id?: string
+      id?: string | null
       sessionID?: string
+      copyChanges?: boolean
     },
     options?: Options<never, ThrowOnError>,
   ) {
@@ -1033,6 +1039,7 @@ export class Workspace extends HeyApiClient {
             { in: "query", key: "workspace" },
             { in: "body", key: "id" },
             { in: "body", key: "sessionID" },
+            { in: "body", key: "copyChanges" },
           ],
         },
       ],
@@ -1554,6 +1561,38 @@ export class Path extends HeyApiClient {
   }
 }
 
+export class Diff extends HeyApiClient {
+  /**
+   * Get raw VCS diff
+   *
+   * Retrieve a raw patch for current uncommitted changes.
+   */
+  public raw<ThrowOnError extends boolean = false>(
+    parameters?: {
+      directory?: string
+      workspace?: string
+    },
+    options?: Options<never, ThrowOnError>,
+  ) {
+    const params = buildClientParams(
+      [parameters],
+      [
+        {
+          args: [
+            { in: "query", key: "directory" },
+            { in: "query", key: "workspace" },
+          ],
+        },
+      ],
+    )
+    return (options?.client ?? this.client).get<VcsDiffRawResponses, unknown, ThrowOnError>({
+      url: "/vcs/diff/raw",
+      ...options,
+      ...params,
+    })
+  }
+}
+
 export class Vcs extends HeyApiClient {
   /**
    * Get VCS info
@@ -1585,6 +1624,36 @@ export class Vcs extends HeyApiClient {
     })
   }
 
+  /**
+   * Get VCS status
+   *
+   * Retrieve changed files in the current working tree without patches.
+   */
+  public status<ThrowOnError extends boolean = false>(
+    parameters?: {
+      directory?: string
+      workspace?: string
+    },
+    options?: Options<never, ThrowOnError>,
+  ) {
+    const params = buildClientParams(
+      [parameters],
+      [
+        {
+          args: [
+            { in: "query", key: "directory" },
+            { in: "query", key: "workspace" },
+          ],
+        },
+      ],
+    )
+    return (options?.client ?? this.client).get<VcsStatusResponses, unknown, ThrowOnError>({
+      url: "/vcs/status",
+      ...options,
+      ...params,
+    })
+  }
+
   /**
    * Get VCS diff
    *
@@ -1616,6 +1685,48 @@ export class Vcs extends HeyApiClient {
       ...params,
     })
   }
+
+  /**
+   * Apply VCS patch
+   *
+   * Apply a raw patch to the current working tree.
+   */
+  public apply<ThrowOnError extends boolean = false>(
+    parameters?: {
+      directory?: string
+      workspace?: string
+      patch?: string
+    },
+    options?: Options<never, ThrowOnError>,
+  ) {
+    const params = buildClientParams(
+      [parameters],
+      [
+        {
+          args: [
+            { in: "query", key: "directory" },
+            { in: "query", key: "workspace" },
+            { in: "body", key: "patch" },
+          ],
+        },
+      ],
+    )
+    return (options?.client ?? this.client).post<VcsApplyResponses, VcsApplyErrors, ThrowOnError>({
+      url: "/vcs/apply",
+      ...options,
+      ...params,
+      headers: {
+        "Content-Type": "application/json",
+        ...options?.headers,
+        ...params.headers,
+      },
+    })
+  }
+
+  private _diff?: Diff
+  get diff2(): Diff {
+    return (this._diff ??= new Diff({ client: this.client }))
+  }
 }
 
 export class Command extends HeyApiClient {
@@ -3320,7 +3431,7 @@ export class Session2 extends HeyApiClient {
         },
       ],
     )
-    return (options?.client ?? this.client).post<SessionForkResponses, unknown, ThrowOnError>({
+    return (options?.client ?? this.client).post<SessionForkResponses, SessionForkErrors, ThrowOnError>({
       url: "/session/{sessionID}/fork",
       ...options,
       ...params,
diff --git a/packages/sdk/js/src/v2/gen/types.gen.ts b/packages/sdk/js/src/v2/gen/types.gen.ts
index 7734ca53eb..175fe69e66 100644
--- a/packages/sdk/js/src/v2/gen/types.gen.ts
+++ b/packages/sdk/js/src/v2/gen/types.gen.ts
@@ -1474,6 +1474,13 @@ export type VcsInfo = {
   default_branch?: string
 }
 
+export type VcsFileStatus = {
+  file: string
+  additions: number
+  deletions: number
+  status: "added" | "deleted" | "modified"
+}
+
 export type VcsFileDiff = {
   file: string
   patch: string
@@ -1482,6 +1489,14 @@ export type VcsFileDiff = {
   status?: "added" | "deleted" | "modified"
 }
 
+export type VcsApplyError = {
+  name: "VcsApplyError"
+  data: {
+    message: string
+    reason: "non-git" | "not-clean"
+  }
+}
+
 export type Command = {
   name: string
   description?: string
@@ -1561,6 +1576,13 @@ export type McpUnsupportedOAuthError = {
   error: string
 }
 
+export type NotFoundError = {
+  name: "NotFoundError"
+  data: {
+    message: string
+  }
+}
+
 export type EffectHttpApiErrorForbidden = {
   _tag: "Forbidden"
 }
@@ -1729,6 +1751,13 @@ export type Workspace = {
   projectID: string
 }
 
+export type WorkspaceWarpError = {
+  name: "WorkspaceWarpError"
+  data: {
+    message: string
+  }
+}
+
 export type SyncEventMessageUpdated = {
   type: "sync"
   name: "message.updated.1"
@@ -3224,13 +3253,6 @@ export type BadRequestError = {
   success: false
 }
 
-export type NotFoundError = {
-  name: "NotFoundError"
-  data: {
-    message: string
-  }
-}
-
 export type AuthRemoveData = {
   body?: never
   path: {
@@ -4020,6 +4042,25 @@ export type VcsGetResponses = {
 
 export type VcsGetResponse = VcsGetResponses[keyof VcsGetResponses]
 
+export type VcsStatusData = {
+  body?: never
+  path?: never
+  query?: {
+    directory?: string
+    workspace?: string
+  }
+  url: "/vcs/status"
+}
+
+export type VcsStatusResponses = {
+  /**
+   * VCS status
+   */
+  200: Array<VcsFileStatus>
+}
+
+export type VcsStatusResponse = VcsStatusResponses[keyof VcsStatusResponses]
+
 export type VcsDiffData = {
   body?: never
   path?: never
@@ -4040,6 +4081,57 @@ export type VcsDiffResponses = {
 
 export type VcsDiffResponse = VcsDiffResponses[keyof VcsDiffResponses]
 
+export type VcsDiffRawData = {
+  body?: never
+  path?: never
+  query?: {
+    directory?: string
+    workspace?: string
+  }
+  url: "/vcs/diff/raw"
+}
+
+export type VcsDiffRawResponses = {
+  /**
+   * Raw VCS diff
+   */
+  200: string
+}
+
+export type VcsDiffRawResponse = VcsDiffRawResponses[keyof VcsDiffRawResponses]
+
+export type VcsApplyData = {
+  body?: {
+    patch: string
+  }
+  path?: never
+  query?: {
+    directory?: string
+    workspace?: string
+  }
+  url: "/vcs/apply"
+}
+
+export type VcsApplyErrors = {
+  /**
+   * VcsApplyError
+   */
+  400: VcsApplyError
+}
+
+export type VcsApplyError2 = VcsApplyErrors[keyof VcsApplyErrors]
+
+export type VcsApplyResponses = {
+  /**
+   * VCS patch applied
+   */
+  200: {
+    applied: boolean
+  }
+}
+
+export type VcsApplyResponse = VcsApplyResponses[keyof VcsApplyResponses]
+
 export type CommandListData = {
   body?: never
   path?: never
@@ -4571,7 +4663,7 @@ export type PtyRemoveData = {
 
 export type PtyRemoveErrors = {
   /**
-   * Not found
+   * NotFoundError
    */
   404: NotFoundError
 }
@@ -4601,7 +4693,7 @@ export type PtyGetData = {
 
 export type PtyGetErrors = {
   /**
-   * Not found
+   * NotFoundError
    */
   404: NotFoundError
 }
@@ -4671,7 +4763,7 @@ export type PtyConnectTokenErrors = {
    */
   403: EffectHttpApiErrorForbidden
   /**
-   * Not found
+   * NotFoundError
    */
   404: NotFoundError
 }
@@ -5070,7 +5162,7 @@ export type SessionDeleteErrors = {
    */
   400: BadRequestError
   /**
-   * Not found
+   * NotFoundError
    */
   404: NotFoundError
 }
@@ -5104,7 +5196,7 @@ export type SessionGetErrors = {
    */
   400: BadRequestError
   /**
-   * Not found
+   * NotFoundError
    */
   404: NotFoundError
 }
@@ -5144,7 +5236,7 @@ export type SessionUpdateErrors = {
    */
   400: BadRequestError
   /**
-   * Not found
+   * NotFoundError
    */
   404: NotFoundError
 }
@@ -5270,7 +5362,7 @@ export type SessionMessagesErrors = {
    */
   400: BadRequestError
   /**
-   * Not found
+   * NotFoundError
    */
   404: NotFoundError
 }
@@ -5395,7 +5487,7 @@ export type SessionMessageErrors = {
    */
   400: BadRequestError
   /**
-   * Not found
+   * NotFoundError
    */
   404: NotFoundError
 }
@@ -5428,6 +5520,15 @@ export type SessionForkData = {
   url: "/session/{sessionID}/fork"
 }
 
+export type SessionForkErrors = {
+  /**
+   * NotFoundError
+   */
+  404: NotFoundError
+}
+
+export type SessionForkError = SessionForkErrors[keyof SessionForkErrors]
+
 export type SessionForkResponses = {
   /**
    * 200
@@ -5527,7 +5628,7 @@ export type SessionUnshareErrors = {
    */
   400: BadRequestError
   /**
-   * Not found
+   * NotFoundError
    */
   404: NotFoundError
 }
@@ -5561,7 +5662,7 @@ export type SessionShareErrors = {
    */
   400: BadRequestError
   /**
-   * Not found
+   * NotFoundError
    */
   404: NotFoundError
 }
@@ -5599,7 +5700,7 @@ export type SessionSummarizeErrors = {
    */
   400: BadRequestError
   /**
-   * Not found
+   * NotFoundError
    */
   404: NotFoundError
 }
@@ -6463,7 +6564,7 @@ export type TuiSelectSessionErrors = {
    */
   400: BadRequestError
   /**
-   * Not found
+   * NotFoundError
    */
   404: NotFoundError
 }
@@ -6656,8 +6757,9 @@ export type ExperimentalWorkspaceRemoveResponse =
 
 export type ExperimentalWorkspaceWarpData = {
   body?: {
-    id: string
+    id: string | null
     sessionID: string
+    copyChanges?: boolean
   }
   path?: never
   query?: {
@@ -6669,9 +6771,9 @@ export type ExperimentalWorkspaceWarpData = {
 
 export type ExperimentalWorkspaceWarpErrors = {
   /**
-   * Bad request
+   * WorkspaceWarpError | VcsApplyError
    */
-  400: BadRequestError
+  400: WorkspaceWarpError | VcsApplyError
 }
 
 export type ExperimentalWorkspaceWarpError = ExperimentalWorkspaceWarpErrors[keyof ExperimentalWorkspaceWarpErrors]
diff --git a/packages/sdk/openapi.json b/packages/sdk/openapi.json
index fea9dd5a95..04c34e2dc1 100644
--- a/packages/sdk/openapi.json
+++ b/packages/sdk/openapi.json
@@ -1897,6 +1897,54 @@
         ]
       }
     },
+    "/vcs/status": {
+      "get": {
+        "tags": ["instance"],
+        "operationId": "vcs.status",
+        "parameters": [
+          {
+            "name": "directory",
+            "in": "query",
+            "required": false,
+            "schema": {
+              "type": "string"
+            }
+          },
+          {
+            "name": "workspace",
+            "in": "query",
+            "required": false,
+            "schema": {
+              "type": "string"
+            }
+          }
+        ],
+        "responses": {
+          "200": {
+            "description": "VCS status",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "type": "array",
+                  "items": {
+                    "$ref": "#/components/schemas/VcsFileStatus"
+                  },
+                  "description": "VCS status"
+                }
+              }
+            }
+          }
+        },
+        "description": "Retrieve changed files in the current working tree without patches.",
+        "summary": "Get VCS status",
+        "x-codeSamples": [
+          {
+            "lang": "js",
+            "source": "import { createOpencodeClient } from \"@opencode-ai/sdk\n\nconst client = createOpencodeClient()\nawait client.vcs.status({\n  ...\n})"
+          }
+        ]
+      }
+    },
     "/vcs/diff": {
       "get": {
         "tags": ["instance"],
@@ -1954,6 +2002,128 @@
         ]
       }
     },
+    "/vcs/diff/raw": {
+      "get": {
+        "tags": ["instance"],
+        "operationId": "vcs.diff.raw",
+        "parameters": [
+          {
+            "name": "directory",
+            "in": "query",
+            "required": false,
+            "schema": {
+              "type": "string"
+            }
+          },
+          {
+            "name": "workspace",
+            "in": "query",
+            "required": false,
+            "schema": {
+              "type": "string"
+            }
+          }
+        ],
+        "responses": {
+          "200": {
+            "description": "Raw VCS diff",
+            "content": {
+              "text/x-diff; charset=utf-8": {
+                "schema": {
+                  "type": "string"
+                }
+              }
+            }
+          }
+        },
+        "description": "Retrieve a raw patch for current uncommitted changes.",
+        "summary": "Get raw VCS diff",
+        "x-codeSamples": [
+          {
+            "lang": "js",
+            "source": "import { createOpencodeClient } from \"@opencode-ai/sdk\n\nconst client = createOpencodeClient()\nawait client.vcs.diff.raw({\n  ...\n})"
+          }
+        ]
+      }
+    },
+    "/vcs/apply": {
+      "post": {
+        "tags": ["instance"],
+        "operationId": "vcs.apply",
+        "parameters": [
+          {
+            "name": "directory",
+            "in": "query",
+            "required": false,
+            "schema": {
+              "type": "string"
+            }
+          },
+          {
+            "name": "workspace",
+            "in": "query",
+            "required": false,
+            "schema": {
+              "type": "string"
+            }
+          }
+        ],
+        "responses": {
+          "200": {
+            "description": "VCS patch applied",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "type": "object",
+                  "properties": {
+                    "applied": {
+                      "type": "boolean"
+                    }
+                  },
+                  "required": ["applied"],
+                  "additionalProperties": false,
+                  "description": "VCS patch applied"
+                }
+              }
+            }
+          },
+          "400": {
+            "description": "VcsApplyError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/VcsApplyError"
+                }
+              }
+            }
+          }
+        },
+        "description": "Apply a raw patch to the current working tree.",
+        "summary": "Apply VCS patch",
+        "requestBody": {
+          "content": {
+            "application/json": {
+              "schema": {
+                "type": "object",
+                "properties": {
+                  "patch": {
+                    "type": "string"
+                  }
+                },
+                "required": ["patch"],
+                "additionalProperties": false
+              }
+            }
+          }
+        },
+        "x-codeSamples": [
+          {
+            "lang": "js",
+            "source": "import { createOpencodeClient } from \"@opencode-ai/sdk\n\nconst client = createOpencodeClient()\nawait client.vcs.apply({\n  ...\n})"
+          }
+        ]
+      }
+    },
     "/command": {
       "get": {
         "tags": ["instance"],
@@ -3241,7 +3411,7 @@
             }
           },
           "404": {
-            "description": "Not found",
+            "description": "NotFoundError",
             "content": {
               "application/json": {
                 "schema": {
@@ -3394,7 +3564,7 @@
             }
           },
           "404": {
-            "description": "Not found",
+            "description": "NotFoundError",
             "content": {
               "application/json": {
                 "schema": {
@@ -3479,7 +3649,7 @@
             }
           },
           "404": {
-            "description": "Not found",
+            "description": "NotFoundError",
             "content": {
               "application/json": {
                 "schema": {
@@ -4454,7 +4624,7 @@
             }
           },
           "404": {
-            "description": "Not found",
+            "description": "NotFoundError",
             "content": {
               "application/json": {
                 "schema": {
@@ -4526,7 +4696,7 @@
             }
           },
           "404": {
-            "description": "Not found",
+            "description": "NotFoundError",
             "content": {
               "application/json": {
                 "schema": {
@@ -4597,7 +4767,7 @@
             }
           },
           "404": {
-            "description": "Not found",
+            "description": "NotFoundError",
             "content": {
               "application/json": {
                 "schema": {
@@ -4952,7 +5122,7 @@
             }
           },
           "404": {
-            "description": "Not found",
+            "description": "NotFoundError",
             "content": {
               "application/json": {
                 "schema": {
@@ -5200,7 +5370,7 @@
             }
           },
           "404": {
-            "description": "Not found",
+            "description": "NotFoundError",
             "content": {
               "application/json": {
                 "schema": {
@@ -5342,6 +5512,16 @@
                 }
               }
             }
+          },
+          "404": {
+            "description": "NotFoundError",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/NotFoundError"
+                }
+              }
+            }
           }
         },
         "description": "Create a new session by forking an existing session at a specific message point.",
@@ -5592,7 +5772,7 @@
             }
           },
           "404": {
-            "description": "Not found",
+            "description": "NotFoundError",
             "content": {
               "application/json": {
                 "schema": {
@@ -5663,7 +5843,7 @@
             }
           },
           "404": {
-            "description": "Not found",
+            "description": "NotFoundError",
             "content": {
               "application/json": {
                 "schema": {
@@ -5737,7 +5917,7 @@
             }
           },
           "404": {
-            "description": "Not found",
+            "description": "NotFoundError",
             "content": {
               "application/json": {
                 "schema": {
@@ -7897,7 +8077,7 @@
             }
           },
           "404": {
-            "description": "Not found",
+            "description": "NotFoundError",
             "content": {
               "application/json": {
                 "schema": {
@@ -8386,11 +8566,18 @@
             "description": "Session warped"
           },
           "400": {
-            "description": "Bad request",
+            "description": "WorkspaceWarpError | VcsApplyError",
             "content": {
               "application/json": {
                 "schema": {
-                  "$ref": "#/components/schemas/BadRequestError"
+                  "anyOf": [
+                    {
+                      "$ref": "#/components/schemas/WorkspaceWarpError"
+                    },
+                    {
+                      "$ref": "#/components/schemas/VcsApplyError"
+                    }
+                  ]
                 }
               }
             }
@@ -8405,10 +8592,20 @@
                 "type": "object",
                 "properties": {
                   "id": {
-                    "type": "string"
+                    "anyOf": [
+                      {
+                        "type": "string"
+                      },
+                      {
+                        "type": "null"
+                      }
+                    ]
                   },
                   "sessionID": {
                     "type": "string"
+                  },
+                  "copyChanges": {
+                    "type": "boolean"
                   }
                 },
                 "required": ["id", "sessionID"],
@@ -12648,6 +12845,28 @@
         },
         "additionalProperties": false
       },
+      "VcsFileStatus": {
+        "type": "object",
+        "properties": {
+          "file": {
+            "type": "string"
+          },
+          "additions": {
+            "type": "integer",
+            "minimum": 0
+          },
+          "deletions": {
+            "type": "integer",
+            "minimum": 0
+          },
+          "status": {
+            "type": "string",
+            "enum": ["added", "deleted", "modified"]
+          }
+        },
+        "required": ["file", "additions", "deletions", "status"],
+        "additionalProperties": false
+      },
       "VcsFileDiff": {
         "type": "object",
         "properties": {
@@ -12673,6 +12892,31 @@
         "required": ["file", "patch", "additions", "deletions"],
         "additionalProperties": false
       },
+      "VcsApplyError": {
+        "type": "object",
+        "properties": {
+          "name": {
+            "type": "string",
+            "enum": ["VcsApplyError"]
+          },
+          "data": {
+            "type": "object",
+            "properties": {
+              "message": {
+                "type": "string"
+              },
+              "reason": {
+                "type": "string",
+                "enum": ["non-git", "not-clean"]
+              }
+            },
+            "required": ["message", "reason"],
+            "additionalProperties": false
+          }
+        },
+        "required": ["name", "data"],
+        "additionalProperties": false
+      },
       "Command": {
         "type": "object",
         "properties": {
@@ -12897,6 +13141,25 @@
         "required": ["error"],
         "additionalProperties": false
       },
+      "NotFoundError": {
+        "type": "object",
+        "required": ["name", "data"],
+        "properties": {
+          "name": {
+            "type": "string",
+            "enum": ["NotFoundError"]
+          },
+          "data": {
+            "type": "object",
+            "required": ["message"],
+            "properties": {
+              "message": {
+                "type": "string"
+              }
+            }
+          }
+        }
+      },
       "effect_HttpApiError_Forbidden": {
         "type": "object",
         "properties": {
@@ -13395,6 +13658,27 @@
         "required": ["id", "type", "name", "branch", "directory", "extra", "projectID"],
         "additionalProperties": false
       },
+      "WorkspaceWarpError": {
+        "type": "object",
+        "properties": {
+          "name": {
+            "type": "string",
+            "enum": ["WorkspaceWarpError"]
+          },
+          "data": {
+            "type": "object",
+            "properties": {
+              "message": {
+                "type": "string"
+              }
+            },
+            "required": ["message"],
+            "additionalProperties": false
+          }
+        },
+        "required": ["name", "data"],
+        "additionalProperties": false
+      },
       "SyncEventMessageUpdated": {
         "type": "object",
         "properties": {
@@ -18019,25 +18303,6 @@
             "enum": [false]
           }
         }
-      },
-      "NotFoundError": {
-        "type": "object",
-        "required": ["name", "data"],
-        "properties": {
-          "name": {
-            "type": "string",
-            "enum": ["NotFoundError"]
-          },
-          "data": {
-            "type": "object",
-            "required": ["message"],
-            "properties": {
-              "message": {
-                "type": "string"
-              }
-            }
-          }
-        }
       }
     }
   },
diff --git a/packages/slack/package.json b/packages/slack/package.json
index f70692d76f..34175d66a2 100644
--- a/packages/slack/package.json
+++ b/packages/slack/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@opencode-ai/slack",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "type": "module",
   "license": "MIT",
   "scripts": {
diff --git a/packages/ui/package.json b/packages/ui/package.json
index f16dfdf134..fc065be9ef 100644
--- a/packages/ui/package.json
+++ b/packages/ui/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@opencode-ai/ui",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "type": "module",
   "license": "MIT",
   "exports": {
diff --git a/packages/ui/src/components/file-ssr.tsx b/packages/ui/src/components/file-ssr.tsx
index ad05555bdf..6f11ca2433 100644
--- a/packages/ui/src/components/file-ssr.tsx
+++ b/packages/ui/src/components/file-ssr.tsx
@@ -128,8 +128,12 @@ function DiffSSRViewer<T>(props: SSRDiffFileProps<T>) {
             prerenderedHTML: local.preloadedDiff.prerenderedHTML,
           }
         : {
-            oldFile: local.before,
-            newFile: local.after,
+            oldFile: local.before
+              ? { ...local.before, contents: typeof local.before.contents === "string" ? local.before.contents : "" }
+              : local.before,
+            newFile: local.after
+              ? { ...local.after, contents: typeof local.after.contents === "string" ? local.after.contents : "" }
+              : local.after,
             lineAnnotations: annotations,
             fileContainer: fileDiffRef,
             containerWrapper: container,
diff --git a/packages/ui/src/components/markdown.tsx b/packages/ui/src/components/markdown.tsx
index 56e2d9d709..7ee73af10f 100644
--- a/packages/ui/src/components/markdown.tsx
+++ b/packages/ui/src/components/markdown.tsx
@@ -33,6 +33,8 @@ const config = {
   SANITIZE_NAMED_PROPS: true,
   FORBID_TAGS: ["style"],
   FORBID_CONTENTS: ["style", "script"],
+  ADD_TAGS: ["svg", "path"],
+  ADD_ATTR: ["d", "viewBox", "preserveAspectRatio", "xmlns"],
 }
 
 const iconPaths = {
diff --git a/packages/ui/src/components/message-part.tsx b/packages/ui/src/components/message-part.tsx
index cc046fdfc5..c36a52f81e 100644
--- a/packages/ui/src/components/message-part.tsx
+++ b/packages/ui/src/components/message-part.tsx
@@ -1906,11 +1906,11 @@ ToolRegistry.register({
                   mode="diff"
                   before={{
                     name: props.metadata?.filediff?.file || props.input.filePath,
-                    contents: props.metadata?.filediff?.before || props.input.oldString,
+                    contents: props.metadata?.filediff?.before || props.input.oldString || "",
                   }}
                   after={{
                     name: props.metadata?.filediff?.file || props.input.filePath,
-                    contents: props.metadata?.filediff?.after || props.input.newString,
+                    contents: props.metadata?.filediff?.after || props.input.newString || "",
                   }}
                 />
               </div>
diff --git a/packages/web/package.json b/packages/web/package.json
index 902e1aa8cd..af6e08a09c 100644
--- a/packages/web/package.json
+++ b/packages/web/package.json
@@ -2,7 +2,7 @@
   "name": "@opencode-ai/web",
   "type": "module",
   "license": "MIT",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "scripts": {
     "dev": "astro dev",
     "dev:remote": "VITE_API_URL=https://api.opencode.ai astro dev",
diff --git a/packages/web/src/content/docs/ar/go.mdx b/packages/web/src/content/docs/ar/go.mdx
index 35c52d9695..81f885335c 100644
--- a/packages/web/src/content/docs/ar/go.mdx
+++ b/packages/web/src/content/docs/ar/go.mdx
@@ -57,10 +57,8 @@ OpenCode Go حاليًا في المرحلة التجريبية.
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -90,10 +88,8 @@ OpenCode Go حاليًا في المرحلة التجريبية.
 | GLM-5              | 1,150               | 2,880              | 5,750            |
 | Kimi K2.5          | 1,850               | 4,630              | 9,250            |
 | Kimi K2.6          | 1,150               | 2,880              | 5,750            |
-| MiMo-V2-Pro        | 1,290               | 3,225              | 6,450            |
-| MiMo-V2-Omni       | 2,150               | 5,450              | 10,900           |
-| MiMo-V2.5-Pro      | 1,290               | 3,225              | 6,450            |
 | MiMo-V2.5 (≤ 256K) | 2,150               | 5,450              | 10,900           |
+| MiMo-V2.5-Pro      | 1,290               | 3,225              | 6,450            |
 | Qwen3.6 Plus       | 3,300               | 8,200              | 16,300           |
 | MiniMax M2.7       | 3,400               | 8,500              | 17,000           |
 | MiniMax M2.5       | 6,300               | 15,900             | 31,800           |
@@ -110,10 +106,8 @@ OpenCode Go حاليًا في المرحلة التجريبية.
 - MiniMax M2.7/M2.5 — ‏300 input، و55,000 cached، و125 output tokens لكل طلب
 - Qwen3.5 Plus — ‏410 input، و47,000 cached، و140 output tokens لكل طلب
 - Qwen3.6 Plus — ‏500 input، و57,000 cached، و190 output tokens لكل طلب
-- MiMo-V2-Pro — ‏350 input، و41,000 cached، و250 output tokens لكل طلب
-- MiMo-V2-Omni — ‏1000 input، و60,000 cached، و140 output tokens لكل طلب
-- MiMo-V2.5-Pro — ‏350 input، و41,000 cached، و250 output tokens لكل طلب
 - MiMo-V2.5 — ‏1000 input، و60,000 cached، و140 output tokens لكل طلب
+- MiMo-V2.5-Pro — ‏350 input، و41,000 cached، و250 output tokens لكل طلب
 
 يمكنك تتبّع استخدامك الحالي في **<a href={console}>console</a>**.
 
@@ -143,10 +137,8 @@ OpenCode Go حاليًا في المرحلة التجريبية.
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/ar/zen.mdx b/packages/web/src/content/docs/ar/zen.mdx
index a2e2aacfe9..33fd9493ba 100644
--- a/packages/web/src/content/docs/ar/zen.mdx
+++ b/packages/web/src/content/docs/ar/zen.mdx
@@ -165,7 +165,7 @@ https://opencode.ai/zen/v1/models
 | GPT 5.1 Codex Mini                | $0.25   | $2.00   | $0.025          | -               |
 | GPT 5                             | $1.07   | $8.50   | $0.107          | -               |
 | GPT 5 Codex                       | $1.07   | $8.50   | $0.107          | -               |
-| GPT 5 Nano                        | Free    | Free    | Free            | -               |
+| GPT 5 Nano                        | $0.05   | $0.40   | $0.005          | -               |
 
 قد تلاحظ _Claude Haiku 3.5_ في سجل الاستخدام. هذا [نموذج منخفض التكلفة](/docs/config/#models) يُستخدم لتوليد عناوين جلساتك.
 
diff --git a/packages/web/src/content/docs/bs/go.mdx b/packages/web/src/content/docs/bs/go.mdx
index a895a20941..d2df6aaad8 100644
--- a/packages/web/src/content/docs/bs/go.mdx
+++ b/packages/web/src/content/docs/bs/go.mdx
@@ -67,10 +67,8 @@ Trenutna lista modela uključuje:
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -100,10 +98,8 @@ Tabela ispod pruža procijenjeni broj zahtjeva na osnovu tipičnih obrazaca kori
 | GLM-5              | 1,150              | 2,880             | 5,750             |
 | Kimi K2.5          | 1,850              | 4,630             | 9,250             |
 | Kimi K2.6          | 1,150              | 2,880             | 5,750             |
-| MiMo-V2-Pro        | 1,290              | 3,225             | 6,450             |
-| MiMo-V2-Omni       | 2,150              | 5,450             | 10,900            |
-| MiMo-V2.5-Pro      | 1,290              | 3,225             | 6,450             |
 | MiMo-V2.5 (≤ 256K) | 2,150              | 5,450             | 10,900            |
+| MiMo-V2.5-Pro      | 1,290              | 3,225             | 6,450             |
 | MiniMax M2.7       | 3,400              | 8,500             | 17,000            |
 | MiniMax M2.5       | 6,300              | 15,900            | 31,800            |
 | Qwen3.6 Plus       | 3,300              | 8,200             | 16,300            |
@@ -120,10 +116,8 @@ Procjene se zasnivaju na zapaženim prosječnim obrascima zahtjeva:
 - MiniMax M2.7/M2.5 — 300 ulaznih, 55,000 keširanih, 125 izlaznih tokena po zahtjevu
 - Qwen3.5 Plus — 410 ulaznih, 47,000 keširanih, 140 izlaznih tokena po zahtjevu
 - Qwen3.6 Plus — 500 ulaznih, 57,000 keširanih, 190 izlaznih tokena po zahtjevu
-- MiMo-V2-Pro — 350 ulaznih, 41,000 keširanih, 250 izlaznih tokena po zahtjevu
-- MiMo-V2-Omni — 1000 ulaznih, 60,000 keširanih, 140 izlaznih tokena po zahtjevu
-- MiMo-V2.5-Pro — 350 ulaznih, 41,000 keširanih, 250 izlaznih tokena po zahtjevu
 - MiMo-V2.5 — 1000 ulaznih, 60,000 keširanih, 140 izlaznih tokena po zahtjevu
+- MiMo-V2.5-Pro — 350 ulaznih, 41,000 keširanih, 250 izlaznih tokena po zahtjevu
 
 Svoju trenutnu potrošnju možete pratiti u **<a href={console}>konzoli</a>**.
 
@@ -155,10 +149,8 @@ Također možete pristupiti Go modelima putem sljedećih API endpointa.
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/bs/zen.mdx b/packages/web/src/content/docs/bs/zen.mdx
index 89527763ca..3723cbaa3c 100644
--- a/packages/web/src/content/docs/bs/zen.mdx
+++ b/packages/web/src/content/docs/bs/zen.mdx
@@ -172,7 +172,7 @@ Podržavamo pay-as-you-go model. Ispod su cijene **po 1M tokena**.
 | GPT 5.1 Codex Mini                | $0.25  | $2.00   | $0.025      | -            |
 | GPT 5                             | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
-| GPT 5 Nano                        | Free   | Free    | Free        | -            |
+| GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |
 
 Možda ćete primijetiti _Claude Haiku 3.5_ u historiji korištenja. To je [low cost model](/docs/config/#models) koji se koristi za generisanje naslova vaših sesija.
 
diff --git a/packages/web/src/content/docs/cli.mdx b/packages/web/src/content/docs/cli.mdx
index 8ecb6a6eb9..ac8a1a3044 100644
--- a/packages/web/src/content/docs/cli.mdx
+++ b/packages/web/src/content/docs/cli.mdx
@@ -61,37 +61,6 @@ opencode agent [command]
 
 ---
 
-### attach
-
-Attach a terminal to an already running OpenCode backend server started via `serve` or `web` commands.
-
-```bash
-opencode attach [url]
-```
-
-This allows using the TUI with a remote OpenCode backend. For example:
-
-```bash
-# Start the backend server for web/mobile access
-opencode web --port 4096 --hostname 0.0.0.0
-
-# In another terminal, attach the TUI to the running backend
-opencode attach http://10.20.30.40:4096
-```
-
-#### Flags
-
-| Flag                                     | Short | Description                                                                |
-| ---------------------------------------- | ----- | -------------------------------------------------------------------------- |
-| <nobr><code>{"--dir"}</code></nobr>      |       | Working directory to start TUI in                                          |
-| <nobr><code>{"--continue"}</code></nobr> | `-c`  | Continue the last session                                                  |
-| <nobr><code>{"--session"}</code></nobr>  | `-s`  | Session ID to continue                                                     |
-| <nobr><code>{"--fork"}</code></nobr>     |       | Fork the session when continuing (use with `--continue` or `--session`)    |
-| <nobr><code>{"--password"}</code></nobr> | `-p`  | Basic auth password (defaults to `OPENCODE_SERVER_PASSWORD`)               |
-| <nobr><code>{"--username"}</code></nobr> | `-u`  | Basic auth username (defaults to `OPENCODE_SERVER_USERNAME` or `opencode`) |
-
----
-
 #### create
 
 Create a new agent with custom configuration.
@@ -126,6 +95,37 @@ opencode agent list
 
 ---
 
+### attach
+
+Attach a terminal to an already running OpenCode backend server started via `serve` or `web` commands.
+
+```bash
+opencode attach [url]
+```
+
+This allows using the TUI with a remote OpenCode backend. For example:
+
+```bash
+# Start the backend server for web/mobile access
+opencode web --port 4096 --hostname 0.0.0.0
+
+# In another terminal, attach the TUI to the running backend
+opencode attach http://10.20.30.40:4096
+```
+
+#### Flags
+
+| Flag                                     | Short | Description                                                                |
+| ---------------------------------------- | ----- | -------------------------------------------------------------------------- |
+| <nobr><code>{"--dir"}</code></nobr>      |       | Working directory to start TUI in                                          |
+| <nobr><code>{"--continue"}</code></nobr> | `-c`  | Continue the last session                                                  |
+| <nobr><code>{"--session"}</code></nobr>  | `-s`  | Session ID to continue                                                     |
+| <nobr><code>{"--fork"}</code></nobr>     |       | Fork the session when continuing (use with `--continue` or `--session`)    |
+| <nobr><code>{"--password"}</code></nobr> | `-p`  | Basic auth password (defaults to `OPENCODE_SERVER_PASSWORD`)               |
+| <nobr><code>{"--username"}</code></nobr> | `-u`  | Basic auth username (defaults to `OPENCODE_SERVER_USERNAME` or `opencode`) |
+
+---
+
 ### auth
 
 Command to manage credentials and login for providers.
diff --git a/packages/web/src/content/docs/da/go.mdx b/packages/web/src/content/docs/da/go.mdx
index db61689a28..6891e6d579 100644
--- a/packages/web/src/content/docs/da/go.mdx
+++ b/packages/web/src/content/docs/da/go.mdx
@@ -67,10 +67,8 @@ Den nuværende liste over modeller inkluderer:
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -100,10 +98,8 @@ Tabellen nedenfor giver et estimeret antal anmodninger baseret på typiske Go-fo
 | GLM-5              | 1,150                   | 2,880               | 5,750                 |
 | Kimi K2.5          | 1,850                   | 4,630               | 9,250                 |
 | Kimi K2.6          | 1,150                   | 2,880               | 5,750                 |
-| MiMo-V2-Pro        | 1,290                   | 3,225               | 6,450                 |
-| MiMo-V2-Omni       | 2,150                   | 5,450               | 10,900                |
-| MiMo-V2.5-Pro      | 1,290                   | 3,225               | 6,450                 |
 | MiMo-V2.5 (≤ 256K) | 2,150                   | 5,450               | 10,900                |
+| MiMo-V2.5-Pro      | 1,290                   | 3,225               | 6,450                 |
 | MiniMax M2.7       | 3,400                   | 8,500               | 17,000                |
 | MiniMax M2.5       | 6,300                   | 15,900              | 31,800                |
 | Qwen3.6 Plus       | 3,300                   | 8,200               | 16,300                |
@@ -120,10 +116,8 @@ Estimaterne er baseret på observerede gennemsnitlige anmodningsmønstre:
 - MiniMax M2.7/M2.5 — 300 input, 55.000 cachelagrede, 125 output-tokens pr. anmodning
 - Qwen3.5 Plus — 410 input, 47.000 cachelagrede, 140 output-tokens pr. anmodning
 - Qwen3.6 Plus — 500 input, 57.000 cachelagrede, 190 output-tokens pr. anmodning
-- MiMo-V2-Pro — 350 input, 41.000 cachelagrede, 250 output-tokens pr. anmodning
-- MiMo-V2-Omni — 1000 input, 60.000 cachelagrede, 140 output-tokens pr. anmodning
-- MiMo-V2.5-Pro — 350 input, 41.000 cachelagrede, 250 output-tokens pr. anmodning
 - MiMo-V2.5 — 1000 input, 60.000 cachelagrede, 140 output-tokens pr. anmodning
+- MiMo-V2.5-Pro — 350 input, 41.000 cachelagrede, 250 output-tokens pr. anmodning
 
 Du kan spore dit nuværende forbrug i **<a href={console}>konsollen</a>**.
 
@@ -155,10 +149,8 @@ Du kan også få adgang til Go-modeller gennem følgende API-endpoints.
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/da/zen.mdx b/packages/web/src/content/docs/da/zen.mdx
index 009ad42023..d45f785a59 100644
--- a/packages/web/src/content/docs/da/zen.mdx
+++ b/packages/web/src/content/docs/da/zen.mdx
@@ -172,7 +172,7 @@ Vi understøtter en pay-as-you-go-model. Nedenfor er priserne **pr. 1M tokens**.
 | GPT 5.1 Codex Mini                | $0.25  | $2.00   | $0.025      | -            |
 | GPT 5                             | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
-| GPT 5 Nano                        | Free   | Free    | Free        | -            |
+| GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |
 
 Du vil måske bemærke _Claude Haiku 3.5_ i din brugshistorik. Det er en [lavprismodel](/docs/config/#models), som bruges til at generere titlerne på dine sessioner.
 
diff --git a/packages/web/src/content/docs/de/go.mdx b/packages/web/src/content/docs/de/go.mdx
index a8da54728d..917ea340ef 100644
--- a/packages/web/src/content/docs/de/go.mdx
+++ b/packages/web/src/content/docs/de/go.mdx
@@ -59,10 +59,8 @@ Die aktuelle Liste der Modelle umfasst:
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -92,10 +90,8 @@ Die folgende Tabelle zeigt eine geschätzte Anzahl von Anfragen basierend auf ty
 | GLM-5              | 1,150                  | 2,880              | 5,750              |
 | Kimi K2.5          | 1,850                  | 4,630              | 9,250              |
 | Kimi K2.6          | 1,150                  | 2,880              | 5,750              |
-| MiMo-V2-Pro        | 1,290                  | 3,225              | 6,450              |
-| MiMo-V2-Omni       | 2,150                  | 5,450              | 10,900             |
-| MiMo-V2.5-Pro      | 1,290                  | 3,225              | 6,450              |
 | MiMo-V2.5 (≤ 256K) | 2,150                  | 5,450              | 10,900             |
+| MiMo-V2.5-Pro      | 1,290                  | 3,225              | 6,450              |
 | MiniMax M2.7       | 3,400                  | 8,500              | 17,000             |
 | MiniMax M2.5       | 6,300                  | 15,900             | 31,800             |
 | Qwen3.6 Plus       | 3,300                  | 8,200              | 16,300             |
@@ -112,10 +108,8 @@ Die Schätzungen basieren auf beobachteten durchschnittlichen Anfragemustern:
 - MiniMax M2.7/M2.5 — 300 Input-, 55.000 Cached-, 125 Output-Tokens pro Anfrage
 - Qwen3.5 Plus — 410 Input-, 47.000 Cached-, 140 Output-Tokens pro Anfrage
 - Qwen3.6 Plus — 500 Input-, 57.000 Cached-, 190 Output-Tokens pro Anfrage
-- MiMo-V2-Pro — 350 Input-, 41.000 Cached-, 250 Output-Tokens pro Anfrage
-- MiMo-V2-Omni — 1.000 Input-, 60.000 Cached-, 140 Output-Tokens pro Anfrage
-- MiMo-V2.5-Pro — 350 Input-, 41.000 Cached-, 250 Output-Tokens pro Anfrage
 - MiMo-V2.5 — 1.000 Input-, 60.000 Cached-, 140 Output-Tokens pro Anfrage
+- MiMo-V2.5-Pro — 350 Input-, 41.000 Cached-, 250 Output-Tokens pro Anfrage
 
 Du kannst deine aktuelle Nutzung in der **<a href={console}>Console</a>** verfolgen.
 
@@ -145,10 +139,8 @@ Du kannst auf die Go-Modelle auch über die folgenden API-Endpunkte zugreifen.
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/de/zen.mdx b/packages/web/src/content/docs/de/zen.mdx
index 11550f61c3..5e6c8eee80 100644
--- a/packages/web/src/content/docs/de/zen.mdx
+++ b/packages/web/src/content/docs/de/zen.mdx
@@ -161,7 +161,7 @@ Wir unterstützen ein Pay-as-you-go-Modell. Unten findest du die Preise **pro 1M
 | GPT 5.1 Codex Mini                | $0.25  | $2.00   | $0.025      | -            |
 | GPT 5                             | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
-| GPT 5 Nano                        | Free   | Free    | Free        | -            |
+| GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |
 
 Möglicherweise siehst du _Claude Haiku 3.5_ in deinem Nutzungsverlauf. Das ist ein [kostengünstiges Modell](/docs/config/#models), das verwendet wird, um die Titel deiner Sessions zu generieren.
 
diff --git a/packages/web/src/content/docs/ecosystem.mdx b/packages/web/src/content/docs/ecosystem.mdx
index 055daf1419..55f0bcdaac 100644
--- a/packages/web/src/content/docs/ecosystem.mdx
+++ b/packages/web/src/content/docs/ecosystem.mdx
@@ -52,6 +52,7 @@ You can also check out [awesome-opencode](https://github.com/awesome-opencode/aw
 | [opencode-worktree](https://github.com/kdcokenny/opencode-worktree)                                | Zero-friction git worktrees for OpenCode                                                          |
 | [opencode-sentry-monitor](https://github.com/stolinski/opencode-sentry-monitor)                    | Trace and debug your AI agents with Sentry AI Monitoring                                          |
 | [opencode-firecrawl](https://github.com/firecrawl/opencode-firecrawl)                              | Web scraping, crawling, and search via the Firecrawl CLI                                          |
+| [opencode-jfrog-plugin](https://github.com/jfrog/opencode-jfrog-plugin)                            | JFrog Plugin for seamless integration of Opencode users to JFrog platform                         |
 
 ---
 
diff --git a/packages/web/src/content/docs/es/go.mdx b/packages/web/src/content/docs/es/go.mdx
index becff7ac04..0be23b3fa4 100644
--- a/packages/web/src/content/docs/es/go.mdx
+++ b/packages/web/src/content/docs/es/go.mdx
@@ -67,10 +67,8 @@ La lista actual de modelos incluye:
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -100,10 +98,8 @@ La siguiente tabla proporciona una cantidad estimada de peticiones basada en los
 | GLM-5              | 1,150                  | 2,880                 | 5,750              |
 | Kimi K2.5          | 1,850                  | 4,630                 | 9,250              |
 | Kimi K2.6          | 1,150                  | 2,880                 | 5,750              |
-| MiMo-V2-Pro        | 1,290                  | 3,225                 | 6,450              |
-| MiMo-V2-Omni       | 2,150                  | 5,450                 | 10,900             |
-| MiMo-V2.5-Pro      | 1,290                  | 3,225                 | 6,450              |
 | MiMo-V2.5 (≤ 256K) | 2,150                  | 5,450                 | 10,900             |
+| MiMo-V2.5-Pro      | 1,290                  | 3,225                 | 6,450              |
 | MiniMax M2.7       | 3,400                  | 8,500                 | 17,000             |
 | MiniMax M2.5       | 6,300                  | 15,900                | 31,800             |
 | Qwen3.6 Plus       | 3,300                  | 8,200                 | 16,300             |
@@ -120,10 +116,8 @@ Las estimaciones se basan en los patrones de peticiones promedio observados:
 - MiniMax M2.7/M2.5 — 300 tokens de entrada, 55,000 en caché, 125 tokens de salida por petición
 - Qwen3.5 Plus — 410 tokens de entrada, 47,000 en caché, 140 tokens de salida por petición
 - Qwen3.6 Plus — 500 tokens de entrada, 57,000 en caché, 190 tokens de salida por petición
-- MiMo-V2-Pro — 350 tokens de entrada, 41,000 en caché, 250 tokens de salida por petición
-- MiMo-V2-Omni — 1000 tokens de entrada, 60,000 en caché, 140 tokens de salida por petición
-- MiMo-V2.5-Pro — 350 tokens de entrada, 41,000 en caché, 250 tokens de salida por petición
 - MiMo-V2.5 — 1000 tokens de entrada, 60,000 en caché, 140 tokens de salida por petición
+- MiMo-V2.5-Pro — 350 tokens de entrada, 41,000 en caché, 250 tokens de salida por petición
 
 Puedes realizar un seguimiento de tu uso actual en la **<a href={console}>consola</a>**.
 
@@ -155,10 +149,8 @@ También puedes acceder a los modelos de Go a través de los siguientes endpoint
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/es/zen.mdx b/packages/web/src/content/docs/es/zen.mdx
index f1a08c7ba5..15436226a5 100644
--- a/packages/web/src/content/docs/es/zen.mdx
+++ b/packages/web/src/content/docs/es/zen.mdx
@@ -172,7 +172,7 @@ Admitimos un modelo de pago por uso. A continuación se muestran los precios **p
 | GPT 5.1 Codex Mini                | $0.25   | $2.00   | $0.025           | -                  |
 | GPT 5                             | $1.07   | $8.50   | $0.107           | -                  |
 | GPT 5 Codex                       | $1.07   | $8.50   | $0.107           | -                  |
-| GPT 5 Nano                        | Free    | Free    | Free             | -                  |
+| GPT 5 Nano                        | $0.05   | $0.40   | $0.005           | -                  |
 
 Puede que notes _Claude Haiku 3.5_ en tu historial de uso. Este es un [modelo de bajo costo](/docs/config/#models) que se usa para generar los títulos de tus sesiones.
 
diff --git a/packages/web/src/content/docs/fr/go.mdx b/packages/web/src/content/docs/fr/go.mdx
index 97280fa372..3dd9c25f32 100644
--- a/packages/web/src/content/docs/fr/go.mdx
+++ b/packages/web/src/content/docs/fr/go.mdx
@@ -57,10 +57,8 @@ La liste actuelle des modèles comprend :
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -90,10 +88,8 @@ Le tableau ci-dessous fournit une estimation du nombre de requêtes basée sur d
 | GLM-5              | 1,150                 | 2,880                | 5,750             |
 | Kimi K2.5          | 1,850                 | 4,630                | 9,250             |
 | Kimi K2.6          | 1,150                 | 2,880                | 5,750             |
-| MiMo-V2-Pro        | 1,290                 | 3,225                | 6,450             |
-| MiMo-V2-Omni       | 2,150                 | 5,450                | 10,900            |
-| MiMo-V2.5-Pro      | 1,290                 | 3,225                | 6,450             |
 | MiMo-V2.5 (≤ 256K) | 2,150                 | 5,450                | 10,900            |
+| MiMo-V2.5-Pro      | 1,290                 | 3,225                | 6,450             |
 | MiniMax M2.7       | 3,400                 | 8,500                | 17,000            |
 | MiniMax M2.5       | 6,300                 | 15,900               | 31,800            |
 | Qwen3.6 Plus       | 3,300                 | 8,200                | 16,300            |
@@ -110,10 +106,8 @@ Les estimations sont basées sur les modèles de requêtes moyens observés :
 - MiniMax M2.7/M2.5 — 300 tokens en entrée, 55,000 en cache, 125 tokens en sortie par requête
 - Qwen3.5 Plus — 410 tokens en entrée, 47,000 en cache, 140 tokens en sortie par requête
 - Qwen3.6 Plus — 500 tokens en entrée, 57,000 en cache, 190 tokens en sortie par requête
-- MiMo-V2-Pro — 350 tokens en entrée, 41,000 en cache, 250 tokens en sortie par requête
-- MiMo-V2-Omni — 1000 tokens en entrée, 60,000 en cache, 140 tokens en sortie par requête
-- MiMo-V2.5-Pro — 350 tokens en entrée, 41,000 en cache, 250 tokens en sortie par requête
 - MiMo-V2.5 — 1000 tokens en entrée, 60,000 en cache, 140 tokens en sortie par requête
+- MiMo-V2.5-Pro — 350 tokens en entrée, 41,000 en cache, 250 tokens en sortie par requête
 
 Vous pouvez suivre votre utilisation actuelle dans la **<a href={console}>console</a>**.
 
@@ -143,10 +137,8 @@ Vous pouvez également accéder aux modèles Go via les points de terminaison d'
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/fr/zen.mdx b/packages/web/src/content/docs/fr/zen.mdx
index 7710da2259..fdf14e8fb0 100644
--- a/packages/web/src/content/docs/fr/zen.mdx
+++ b/packages/web/src/content/docs/fr/zen.mdx
@@ -161,7 +161,7 @@ Nous prenons en charge un modèle de paiement à l'utilisation. Vous trouverez c
 | GPT 5.1 Codex Mini                | $0.25  | $2.00   | $0.025      | -            |
 | GPT 5                             | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
-| GPT 5 Nano                        | Free   | Free    | Free        | -            |
+| GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |
 
 Vous remarquerez peut-être _Claude Haiku 3.5_ dans votre historique d'utilisation. Il s'agit d'un [modèle à faible coût](/docs/config/#models) utilisé pour générer les titres de vos sessions.
 
diff --git a/packages/web/src/content/docs/go.mdx b/packages/web/src/content/docs/go.mdx
index cddb6d491b..237d1c4b84 100644
--- a/packages/web/src/content/docs/go.mdx
+++ b/packages/web/src/content/docs/go.mdx
@@ -67,10 +67,8 @@ The current list of models includes:
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **MiniMax M2.7**
 - **Qwen3.5 Plus**
@@ -100,10 +98,8 @@ The table below provides an estimated request count based on typical Go usage pa
 | GLM-5              | 1,150               | 2,880             | 5,750              |
 | Kimi K2.5          | 1,850               | 4,630             | 9,250              |
 | Kimi K2.6          | 1,150               | 2,880             | 5,750              |
-| MiMo-V2-Pro        | 1,290               | 3,225             | 6,450              |
-| MiMo-V2-Omni       | 2,150               | 5,450             | 10,900             |
-| MiMo-V2.5-Pro      | 1,290               | 3,225             | 6,450              |
 | MiMo-V2.5 (≤ 256K) | 2,150               | 5,450             | 10,900             |
+| MiMo-V2.5-Pro      | 1,290               | 3,225             | 6,450              |
 | MiniMax M2.7       | 3,400               | 8,500             | 17,000             |
 | MiniMax M2.5       | 6,300               | 15,900            | 31,800             |
 | Qwen3.6 Plus       | 3,300               | 8,200             | 16,300             |
@@ -118,10 +114,8 @@ Estimates are based on observed average request patterns:
 - DeepSeek V4 Pro — 750 input, 82,000 cached, 290 output tokens per request
 - DeepSeek V4 Flash — 790 input, 68,000 cached, 280 output tokens per request
 - MiniMax M2.7/M2.5 — 300 input, 55,000 cached, 125 output tokens per request
-- MiMo-V2-Pro — 350 input, 41,000 cached, 250 output tokens per request
-- MiMo-V2-Omni — 1000 input, 60,000 cached, 140 output tokens per request
-- MiMo-V2.5-Pro — 350 input, 41,000 cached, 250 output tokens per request
 - MiMo-V2.5 — 1000 input, 60,000 cached, 140 output tokens per request
+- MiMo-V2.5-Pro — 350 input, 41,000 cached, 250 output tokens per request
 - Qwen3.5 Plus — 410 input, 47,000 cached, 140 output tokens per request
 - Qwen3.6 Plus — 500 input, 57,000 cached, 190 output tokens per request
 
@@ -155,10 +149,8 @@ You can also access Go models through the following API endpoints.
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/it/go.mdx b/packages/web/src/content/docs/it/go.mdx
index 28f8c5fbf8..df4f6dd1ca 100644
--- a/packages/web/src/content/docs/it/go.mdx
+++ b/packages/web/src/content/docs/it/go.mdx
@@ -65,10 +65,8 @@ L'elenco attuale dei modelli include:
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -98,10 +96,8 @@ La tabella seguente fornisce una stima del conteggio delle richieste in base a p
 | GLM-5              | 1,150                | 2,880                 | 5,750             |
 | Kimi K2.5          | 1,850                | 4,630                 | 9,250             |
 | Kimi K2.6          | 1,150                | 2,880                 | 5,750             |
-| MiMo-V2-Pro        | 1,290                | 3,225                 | 6,450             |
-| MiMo-V2-Omni       | 2,150                | 5,450                 | 10,900            |
-| MiMo-V2.5-Pro      | 1,290                | 3,225                 | 6,450             |
 | MiMo-V2.5 (≤ 256K) | 2,150                | 5,450                 | 10,900            |
+| MiMo-V2.5-Pro      | 1,290                | 3,225                 | 6,450             |
 | MiniMax M2.7       | 3,400                | 8,500                 | 17,000            |
 | MiniMax M2.5       | 6,300                | 15,900                | 31,800            |
 | Qwen3.6 Plus       | 3,300                | 8,200                 | 16,300            |
@@ -118,10 +114,8 @@ Le stime si basano sui pattern medi di richieste osservati:
 - MiniMax M2.7/M2.5 — 300 di input, 55.000 in cache, 125 token di output per richiesta
 - Qwen3.5 Plus — 410 di input, 47.000 in cache, 140 token di output per richiesta
 - Qwen3.6 Plus — 500 di input, 57.000 in cache, 190 token di output per richiesta
-- MiMo-V2-Pro — 350 di input, 41.000 in cache, 250 token di output per richiesta
-- MiMo-V2-Omni — 1000 di input, 60.000 in cache, 140 token di output per richiesta
-- MiMo-V2.5-Pro — 350 di input, 41.000 in cache, 250 token di output per richiesta
 - MiMo-V2.5 — 1000 di input, 60.000 in cache, 140 token di output per richiesta
+- MiMo-V2.5-Pro — 350 di input, 41.000 in cache, 250 token di output per richiesta
 
 Puoi monitorare il tuo utilizzo attuale nella **<a href={console}>console</a>**.
 
@@ -153,10 +147,8 @@ Puoi anche accedere ai modelli Go tramite i seguenti endpoint API.
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/it/zen.mdx b/packages/web/src/content/docs/it/zen.mdx
index a3b8725535..a53d6a2ba1 100644
--- a/packages/web/src/content/docs/it/zen.mdx
+++ b/packages/web/src/content/docs/it/zen.mdx
@@ -172,7 +172,7 @@ Supportiamo un modello pay-as-you-go. Qui sotto trovi i prezzi **per 1M token**.
 | GPT 5.1 Codex Mini                | $0.25  | $2.00   | $0.025      | -            |
 | GPT 5                             | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
-| GPT 5 Nano                        | Free   | Free    | Free        | -            |
+| GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |
 
 Potresti notare _Claude Haiku 3.5_ nella cronologia di utilizzo. È un [modello a basso costo](/docs/config/#models) usato per generare i titoli delle tue sessioni.
 
diff --git a/packages/web/src/content/docs/ja/go.mdx b/packages/web/src/content/docs/ja/go.mdx
index 5f4fcfbc39..0cb294754f 100644
--- a/packages/web/src/content/docs/ja/go.mdx
+++ b/packages/web/src/content/docs/ja/go.mdx
@@ -57,10 +57,8 @@ OpenCode Goをサブスクライブできるのは、1つのワークスペー
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -90,10 +88,8 @@ OpenCode Goには以下の制限が含まれています：
 | GLM-5              | 1,150                     | 2,880            | 5,750            |
 | Kimi K2.5          | 1,850                     | 4,630            | 9,250            |
 | Kimi K2.6          | 1,150                     | 2,880            | 5,750            |
-| MiMo-V2-Pro        | 1,290                     | 3,225            | 6,450            |
-| MiMo-V2-Omni       | 2,150                     | 5,450            | 10,900           |
-| MiMo-V2.5-Pro      | 1,290                     | 3,225            | 6,450            |
 | MiMo-V2.5 (≤ 256K) | 2,150                     | 5,450            | 10,900           |
+| MiMo-V2.5-Pro      | 1,290                     | 3,225            | 6,450            |
 | MiniMax M2.7       | 3,400                     | 8,500            | 17,000           |
 | MiniMax M2.5       | 6,300                     | 15,900           | 31,800           |
 | Qwen3.6 Plus       | 3,300                     | 8,200            | 16,300           |
@@ -110,10 +106,8 @@ OpenCode Goには以下の制限が含まれています：
 - MiniMax M2.7/M2.5 — リクエストあたり 入力 300トークン、キャッシュ 55,000トークン、出力 125トークン
 - Qwen3.5 Plus — リクエストあたり 入力 410トークン、キャッシュ 47,000トークン、出力 140トークン
 - Qwen3.6 Plus — リクエストあたり 入力 500トークン、キャッシュ 57,000トークン、出力 190トークン
-- MiMo-V2-Pro — リクエストあたり 入力 350トークン、キャッシュ 41,000トークン、出力 250トークン
-- MiMo-V2-Omni — リクエストあたり 入力 1000トークン、キャッシュ 60,000トークン、出力 140トークン
-- MiMo-V2.5-Pro — リクエストあたり 入力 350トークン、キャッシュ 41,000トークン、出力 250トークン
 - MiMo-V2.5 — リクエストあたり 入力 1000トークン、キャッシュ 60,000トークン、出力 140トークン
+- MiMo-V2.5-Pro — リクエストあたり 入力 350トークン、キャッシュ 41,000トークン、出力 250トークン
 
 現在の利用状況は**<a href={console}>コンソール</a>**で追跡できます。
 
@@ -143,10 +137,8 @@ Zen残高にクレジットがある場合は、コンソールで**Use balance*
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/ja/zen.mdx b/packages/web/src/content/docs/ja/zen.mdx
index 8fcdc6d46b..64427a72ec 100644
--- a/packages/web/src/content/docs/ja/zen.mdx
+++ b/packages/web/src/content/docs/ja/zen.mdx
@@ -161,7 +161,7 @@ https://opencode.ai/zen/v1/models
 | GPT 5.1 Codex Mini                | $0.25  | $2.00   | $0.025      | -            |
 | GPT 5                             | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
-| GPT 5 Nano                        | Free   | Free    | Free        | -            |
+| GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |
 
 使用履歴に _Claude Haiku 3.5_ が表示されることがあります。これはセッションのタイトル生成に使われる [low cost model](/docs/config/#models) です。
 
diff --git a/packages/web/src/content/docs/ko/go.mdx b/packages/web/src/content/docs/ko/go.mdx
index ef05b01c49..d0a3b9d0d1 100644
--- a/packages/web/src/content/docs/ko/go.mdx
+++ b/packages/web/src/content/docs/ko/go.mdx
@@ -57,10 +57,8 @@ workspace당 한 명의 멤버만 OpenCode Go를 구독할 수 있습니다.
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -90,10 +88,8 @@ OpenCode Go에는 다음과 같은 한도가 포함됩니다.
 | GLM-5              | 1,150             | 2,880          | 5,750          |
 | Kimi K2.5          | 1,850             | 4,630          | 9,250          |
 | Kimi K2.6          | 1,150             | 2,880          | 5,750          |
-| MiMo-V2-Pro        | 1,290             | 3,225          | 6,450          |
-| MiMo-V2-Omni       | 2,150             | 5,450          | 10,900         |
-| MiMo-V2.5-Pro      | 1,290             | 3,225          | 6,450          |
 | MiMo-V2.5 (≤ 256K) | 2,150             | 5,450          | 10,900         |
+| MiMo-V2.5-Pro      | 1,290             | 3,225          | 6,450          |
 | MiniMax M2.7       | 3,400             | 8,500          | 17,000         |
 | MiniMax M2.5       | 6,300             | 15,900         | 31,800         |
 | Qwen3.6 Plus       | 3,300             | 8,200          | 16,300         |
@@ -110,10 +106,8 @@ OpenCode Go에는 다음과 같은 한도가 포함됩니다.
 - MiniMax M2.7/M2.5 — 요청당 입력 300, 캐시 55,000, 출력 토큰 125
 - Qwen3.5 Plus — 요청당 입력 410, 캐시 47,000, 출력 토큰 140
 - Qwen3.6 Plus — 요청당 입력 500, 캐시 57,000, 출력 토큰 190
-- MiMo-V2-Pro — 요청당 입력 350, 캐시 41,000, 출력 토큰 250
-- MiMo-V2-Omni — 요청당 입력 1000, 캐시 60,000, 출력 토큰 140
-- MiMo-V2.5-Pro — 요청당 입력 350, 캐시 41,000, 출력 토큰 250
 - MiMo-V2.5 — 요청당 입력 1000, 캐시 60,000, 출력 토큰 140
+- MiMo-V2.5-Pro — 요청당 입력 350, 캐시 41,000, 출력 토큰 250
 
 현재 사용량은 **<a href={console}>console</a>**에서 확인할 수 있습니다.
 
@@ -143,10 +137,8 @@ Zen 잔액에 크레딧도 있다면, console에서 **Use balance** 옵션을 
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/ko/zen.mdx b/packages/web/src/content/docs/ko/zen.mdx
index eb99c29fe6..e80a5e8710 100644
--- a/packages/web/src/content/docs/ko/zen.mdx
+++ b/packages/web/src/content/docs/ko/zen.mdx
@@ -161,7 +161,7 @@ https://opencode.ai/zen/v1/models
 | GPT 5.1 Codex Mini                | $0.25  | $2.00   | $0.025      | -            |
 | GPT 5                             | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
-| GPT 5 Nano                        | Free   | Free    | Free        | -            |
+| GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |
 
 사용 기록에서 *Claude Haiku 3.5*를 볼 수 있습니다. 이는 세션 제목을 생성할 때 사용되는 [저비용 모델](/docs/config/#models)입니다.
 
diff --git a/packages/web/src/content/docs/nb/go.mdx b/packages/web/src/content/docs/nb/go.mdx
index 02a2ba9e0b..e19b6ccce1 100644
--- a/packages/web/src/content/docs/nb/go.mdx
+++ b/packages/web/src/content/docs/nb/go.mdx
@@ -67,10 +67,8 @@ Den nåværende listen over modeller inkluderer:
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -100,10 +98,8 @@ Tabellen nedenfor gir et estimert antall forespørsler basert på typiske bruksm
 | GLM-5              | 1,150                    | 2,880                | 5,750                  |
 | Kimi K2.5          | 1,850                    | 4,630                | 9,250                  |
 | Kimi K2.6          | 1,150                    | 2,880                | 5,750                  |
-| MiMo-V2-Pro        | 1,290                    | 3,225                | 6,450                  |
-| MiMo-V2-Omni       | 2,150                    | 5,450                | 10,900                 |
-| MiMo-V2.5-Pro      | 1,290                    | 3,225                | 6,450                  |
 | MiMo-V2.5 (≤ 256K) | 2,150                    | 5,450                | 10,900                 |
+| MiMo-V2.5-Pro      | 1,290                    | 3,225                | 6,450                  |
 | MiniMax M2.7       | 3,400                    | 8,500                | 17,000                 |
 | MiniMax M2.5       | 6,300                    | 15,900               | 31,800                 |
 | Qwen3.6 Plus       | 3,300                    | 8,200                | 16,300                 |
@@ -120,10 +116,8 @@ Estimatene er basert på observerte gjennomsnittlige forespørselsmønstre:
 - MiniMax M2.7/M2.5 — 300 input, 55 000 bufret, 125 output-tokens per forespørsel
 - Qwen3.5 Plus — 410 input, 47 000 bufret, 140 output-tokens per forespørsel
 - Qwen3.6 Plus — 500 input, 57 000 bufret, 190 output-tokens per forespørsel
-- MiMo-V2-Pro — 350 input, 41 000 bufret, 250 output-tokens per forespørsel
-- MiMo-V2-Omni — 1000 input, 60 000 bufret, 140 output-tokens per forespørsel
-- MiMo-V2.5-Pro — 350 input, 41 000 bufret, 250 output-tokens per forespørsel
 - MiMo-V2.5 — 1000 input, 60 000 bufret, 140 output-tokens per forespørsel
+- MiMo-V2.5-Pro — 350 input, 41 000 bufret, 250 output-tokens per forespørsel
 
 Du kan spore din nåværende bruk i **<a href={console}>konsollen</a>**.
 
@@ -155,10 +149,8 @@ Du kan også få tilgang til Go-modeller gjennom følgende API-endepunkter.
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/nb/zen.mdx b/packages/web/src/content/docs/nb/zen.mdx
index 8ab1762e1f..4bd1e6115e 100644
--- a/packages/web/src/content/docs/nb/zen.mdx
+++ b/packages/web/src/content/docs/nb/zen.mdx
@@ -172,7 +172,7 @@ Vi støtter en pay-as-you-go-modell. Nedenfor er prisene **per 1M tokens**.
 | GPT 5.1 Codex Mini                | $0.25   | $2.00   | $0.025        | -               |
 | GPT 5                             | $1.07   | $8.50   | $0.107        | -               |
 | GPT 5 Codex                       | $1.07   | $8.50   | $0.107        | -               |
-| GPT 5 Nano                        | Free    | Free    | Free          | -               |
+| GPT 5 Nano                        | $0.05   | $0.40   | $0.005        | -               |
 
 Du vil kanskje legge merke til _Claude Haiku 3.5_ i brukshistorikken din. Dette er en [lavprismodell](/docs/config/#models) som brukes til å generere titlene på øktene dine.
 
diff --git a/packages/web/src/content/docs/pl/go.mdx b/packages/web/src/content/docs/pl/go.mdx
index 224671a19a..00f76a103f 100644
--- a/packages/web/src/content/docs/pl/go.mdx
+++ b/packages/web/src/content/docs/pl/go.mdx
@@ -61,10 +61,8 @@ Obecna lista modeli obejmuje:
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -94,10 +92,8 @@ Poniższa tabela przedstawia szacunkową liczbę żądań na podstawie typowych
 | GLM-5              | 1,150               | 2,880              | 5,750              |
 | Kimi K2.5          | 1,850               | 4,630              | 9,250              |
 | Kimi K2.6          | 1,150               | 2,880              | 5,750              |
-| MiMo-V2-Pro        | 1,290               | 3,225              | 6,450              |
-| MiMo-V2-Omni       | 2,150               | 5,450              | 10,900             |
-| MiMo-V2.5-Pro      | 1,290               | 3,225              | 6,450              |
 | MiMo-V2.5 (≤ 256K) | 2,150               | 5,450              | 10,900             |
+| MiMo-V2.5-Pro      | 1,290               | 3,225              | 6,450              |
 | MiniMax M2.7       | 3,400               | 8,500              | 17,000             |
 | MiniMax M2.5       | 6,300               | 15,900             | 31,800             |
 | Qwen3.6 Plus       | 3,300               | 8,200              | 16,300             |
@@ -114,10 +110,8 @@ Szacunki opierają się na zaobserwowanych średnich wzorcach żądań:
 - MiniMax M2.7/M2.5 — 300 tokenów wejściowych, 55 000 w pamięci podręcznej, 125 tokenów wyjściowych na żądanie
 - Qwen3.5 Plus — 410 tokenów wejściowych, 47 000 w pamięci podręcznej, 140 tokenów wyjściowych na żądanie
 - Qwen3.6 Plus — 500 tokenów wejściowych, 57 000 w pamięci podręcznej, 190 tokenów wyjściowych na żądanie
-- MiMo-V2-Pro — 350 tokenów wejściowych, 41 000 w pamięci podręcznej, 250 tokenów wyjściowych na żądanie
-- MiMo-V2-Omni — 1000 tokenów wejściowych, 60 000 w pamięci podręcznej, 140 tokenów wyjściowych na żądanie
-- MiMo-V2.5-Pro — 350 tokenów wejściowych, 41 000 w pamięci podręcznej, 250 tokenów wyjściowych na żądanie
 - MiMo-V2.5 — 1000 tokenów wejściowych, 60 000 w pamięci podręcznej, 140 tokenów wyjściowych na żądanie
+- MiMo-V2.5-Pro — 350 tokenów wejściowych, 41 000 w pamięci podręcznej, 250 tokenów wyjściowych na żądanie
 
 Możesz śledzić swoje bieżące zużycie w **<a href={console}>konsoli</a>**.
 
@@ -147,10 +141,8 @@ Możesz również uzyskać dostęp do modeli Go za pośrednictwem następującyc
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/pl/zen.mdx b/packages/web/src/content/docs/pl/zen.mdx
index 52906036c0..ebd16d7856 100644
--- a/packages/web/src/content/docs/pl/zen.mdx
+++ b/packages/web/src/content/docs/pl/zen.mdx
@@ -172,7 +172,7 @@ Obsługujemy model pay-as-you-go. Poniżej znajdują się ceny **za 1M tokenów*
 | GPT 5.1 Codex Mini                | $0.25   | $2.00   | $0.025         | -              |
 | GPT 5                             | $1.07   | $8.50   | $0.107         | -              |
 | GPT 5 Codex                       | $1.07   | $8.50   | $0.107         | -              |
-| GPT 5 Nano                        | Free    | Free    | Free           | -              |
+| GPT 5 Nano                        | $0.05   | $0.40   | $0.005         | -              |
 
 Możesz zauważyć _Claude Haiku 3.5_ w historii użycia. To [niedrogi model](/docs/config/#models),
 który służy do generowania tytułów Twoich sesji.
diff --git a/packages/web/src/content/docs/pt-br/go.mdx b/packages/web/src/content/docs/pt-br/go.mdx
index e50f7d3962..44c5092a00 100644
--- a/packages/web/src/content/docs/pt-br/go.mdx
+++ b/packages/web/src/content/docs/pt-br/go.mdx
@@ -67,10 +67,8 @@ A lista atual de modelos inclui:
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -100,10 +98,8 @@ A tabela abaixo fornece uma contagem estimada de requisições com base nos padr
 | GLM-5              | 1,150                   | 2,880                  | 5,750               |
 | Kimi K2.5          | 1,850                   | 4,630                  | 9,250               |
 | Kimi K2.6          | 1,150                   | 2,880                  | 5,750               |
-| MiMo-V2-Pro        | 1,290                   | 3,225                  | 6,450               |
-| MiMo-V2-Omni       | 2,150                   | 5,450                  | 10,900              |
-| MiMo-V2.5-Pro      | 1,290                   | 3,225                  | 6,450               |
 | MiMo-V2.5 (≤ 256K) | 2,150                   | 5,450                  | 10,900              |
+| MiMo-V2.5-Pro      | 1,290                   | 3,225                  | 6,450               |
 | MiniMax M2.7       | 3,400                   | 8,500                  | 17,000              |
 | MiniMax M2.5       | 6,300                   | 15,900                 | 31,800              |
 | Qwen3.6 Plus       | 3,300                   | 8,200                  | 16,300              |
@@ -120,10 +116,8 @@ As estimativas baseiam-se nos padrões médios de requisições observados:
 - MiniMax M2.7/M2.5 — 300 tokens de entrada, 55.000 em cache, 125 tokens de saída por requisição
 - Qwen3.5 Plus — 410 tokens de entrada, 47.000 em cache, 140 tokens de saída por requisição
 - Qwen3.6 Plus — 500 tokens de entrada, 57.000 em cache, 190 tokens de saída por requisição
-- MiMo-V2-Pro — 350 tokens de entrada, 41.000 em cache, 250 tokens de saída por requisição
-- MiMo-V2-Omni — 1000 tokens de entrada, 60.000 em cache, 140 tokens de saída por requisição
-- MiMo-V2.5-Pro — 350 tokens de entrada, 41.000 em cache, 250 tokens de saída por requisição
 - MiMo-V2.5 — 1000 tokens de entrada, 60.000 em cache, 140 tokens de saída por requisição
+- MiMo-V2.5-Pro — 350 tokens de entrada, 41.000 em cache, 250 tokens de saída por requisição
 
 Você pode acompanhar o seu uso atual no **<a href={console}>console</a>**.
 
@@ -155,10 +149,8 @@ Você também pode acessar os modelos do Go através dos seguintes endpoints de
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/pt-br/zen.mdx b/packages/web/src/content/docs/pt-br/zen.mdx
index b35cfdbde5..1dcc98c5d5 100644
--- a/packages/web/src/content/docs/pt-br/zen.mdx
+++ b/packages/web/src/content/docs/pt-br/zen.mdx
@@ -161,7 +161,7 @@ Oferecemos um modelo pay-as-you-go. Abaixo estão os preços **por 1M tokens**.
 | GPT 5.1 Codex Mini                | $0.25   | $2.00   | $0.025           | -                |
 | GPT 5                             | $1.07   | $8.50   | $0.107           | -                |
 | GPT 5 Codex                       | $1.07   | $8.50   | $0.107           | -                |
-| GPT 5 Nano                        | Free    | Free    | Free             | -                |
+| GPT 5 Nano                        | $0.05   | $0.40   | $0.005           | -                |
 
 Você pode notar _Claude Haiku 3.5_ no seu histórico de uso. Este é um [low cost model](/docs/config/#models) usado para gerar os títulos das suas sessões.
 
diff --git a/packages/web/src/content/docs/ru/go.mdx b/packages/web/src/content/docs/ru/go.mdx
index 4f11204e1a..66e929c5f4 100644
--- a/packages/web/src/content/docs/ru/go.mdx
+++ b/packages/web/src/content/docs/ru/go.mdx
@@ -67,10 +67,8 @@ OpenCode Go работает так же, как и любой другой пр
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -100,10 +98,8 @@ OpenCode Go включает следующие лимиты:
 | GLM-5              | 1,150               | 2,880             | 5,750            |
 | Kimi K2.5          | 1,850               | 4,630             | 9,250            |
 | Kimi K2.6          | 1,150               | 2,880             | 5,750            |
-| MiMo-V2-Pro        | 1,290               | 3,225             | 6,450            |
-| MiMo-V2-Omni       | 2,150               | 5,450             | 10,900           |
-| MiMo-V2.5-Pro      | 1,290               | 3,225             | 6,450            |
 | MiMo-V2.5 (≤ 256K) | 2,150               | 5,450             | 10,900           |
+| MiMo-V2.5-Pro      | 1,290               | 3,225             | 6,450            |
 | MiniMax M2.7       | 3,400               | 8,500             | 17,000           |
 | MiniMax M2.5       | 6,300               | 15,900            | 31,800           |
 | Qwen3.6 Plus       | 3,300               | 8,200             | 16,300           |
@@ -120,10 +116,8 @@ OpenCode Go включает следующие лимиты:
 - MiniMax M2.7/M2.5 — 300 входных, 55,000 кешированных, 125 выходных токенов на запрос
 - Qwen3.5 Plus — 410 входных, 47,000 кешированных, 140 выходных токенов на запрос
 - Qwen3.6 Plus — 500 входных, 57,000 кешированных, 190 выходных токенов на запрос
-- MiMo-V2-Pro — 350 входных, 41,000 кешированных, 250 выходных токенов на запрос
-- MiMo-V2-Omni — 1000 входных, 60,000 кешированных, 140 выходных токенов на запрос
-- MiMo-V2.5-Pro — 350 входных, 41,000 кешированных, 250 выходных токенов на запрос
 - MiMo-V2.5 — 1000 входных, 60,000 кешированных, 140 выходных токенов на запрос
+- MiMo-V2.5-Pro — 350 входных, 41,000 кешированных, 250 выходных токенов на запрос
 
 Вы можете отслеживать текущее использование в **<a href={console}>консоли</a>**.
 
@@ -155,10 +149,8 @@ OpenCode Go включает следующие лимиты:
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/ru/zen.mdx b/packages/web/src/content/docs/ru/zen.mdx
index 919026447f..10c55fc4dd 100644
--- a/packages/web/src/content/docs/ru/zen.mdx
+++ b/packages/web/src/content/docs/ru/zen.mdx
@@ -172,7 +172,7 @@ https://opencode.ai/zen/v1/models
 | GPT 5.1 Codex Mini                | $0.25  | $2.00   | $0.025      | -            |
 | GPT 5                             | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
-| GPT 5 Nano                        | Free   | Free    | Free        | -            |
+| GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |
 
 Вы можете заметить _Claude Haiku 3.5_ в истории использования. Это [недорогая модель](/docs/config/#models), которая используется для генерации заголовков ваших сессий.
 
diff --git a/packages/web/src/content/docs/th/go.mdx b/packages/web/src/content/docs/th/go.mdx
index 2a4c90a840..1fa0f8cc2a 100644
--- a/packages/web/src/content/docs/th/go.mdx
+++ b/packages/web/src/content/docs/th/go.mdx
@@ -57,10 +57,8 @@ OpenCode Go ทำงานเหมือนกับผู้ให้บร
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -90,10 +88,8 @@ OpenCode Go มีขีดจำกัดดังต่อไปนี้:
 | GLM-5              | 1,150                  | 2,880               | 5,750             |
 | Kimi K2.5          | 1,850                  | 4,630               | 9,250             |
 | Kimi K2.6          | 1,150                  | 2,880               | 5,750             |
-| MiMo-V2-Pro        | 1,290                  | 3,225               | 6,450             |
-| MiMo-V2-Omni       | 2,150                  | 5,450               | 10,900            |
-| MiMo-V2.5-Pro      | 1,290                  | 3,225               | 6,450             |
 | MiMo-V2.5 (≤ 256K) | 2,150                  | 5,450               | 10,900            |
+| MiMo-V2.5-Pro      | 1,290                  | 3,225               | 6,450             |
 | MiniMax M2.7       | 3,400                  | 8,500               | 17,000            |
 | MiniMax M2.5       | 6,300                  | 15,900              | 31,800            |
 | Qwen3.6 Plus       | 3,300                  | 8,200               | 16,300            |
@@ -110,10 +106,8 @@ OpenCode Go มีขีดจำกัดดังต่อไปนี้:
 - MiniMax M2.7/M2.5 — 300 input, 55,000 cached, 125 output tokens ต่อ request
 - Qwen3.5 Plus — 410 input, 47,000 cached, 140 output tokens ต่อ request
 - Qwen3.6 Plus — 500 input, 57,000 cached, 190 output tokens ต่อ request
-- MiMo-V2-Pro — 350 input, 41,000 cached, 250 output tokens ต่อ request
-- MiMo-V2-Omni — 1000 input, 60,000 cached, 140 output tokens ต่อ request
-- MiMo-V2.5-Pro — 350 input, 41,000 cached, 250 output tokens ต่อ request
 - MiMo-V2.5 — 1000 input, 60,000 cached, 140 output tokens ต่อ request
+- MiMo-V2.5-Pro — 350 input, 41,000 cached, 250 output tokens ต่อ request
 
 คุณสามารถติดตามการใช้งานปัจจุบันของคุณได้ใน **<a href={console}>console</a>**
 
@@ -143,10 +137,8 @@ OpenCode Go มีขีดจำกัดดังต่อไปนี้:
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/th/zen.mdx b/packages/web/src/content/docs/th/zen.mdx
index b3914b73c2..cb2556ef63 100644
--- a/packages/web/src/content/docs/th/zen.mdx
+++ b/packages/web/src/content/docs/th/zen.mdx
@@ -163,7 +163,7 @@ https://opencode.ai/zen/v1/models
 | GPT 5.1 Codex Mini                | $0.25  | $2.00   | $0.025      | -            |
 | GPT 5                             | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
-| GPT 5 Nano                        | Free   | Free    | Free        | -            |
+| GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |
 
 คุณอาจสังเกตเห็น _Claude Haiku 3.5_ ในประวัติการใช้งานของคุณ นี่คือ [low cost model](/docs/config/#models) ที่ใช้สร้างชื่อ session ของคุณ
 
diff --git a/packages/web/src/content/docs/tr/go.mdx b/packages/web/src/content/docs/tr/go.mdx
index b3995e8a57..367be5a750 100644
--- a/packages/web/src/content/docs/tr/go.mdx
+++ b/packages/web/src/content/docs/tr/go.mdx
@@ -57,10 +57,8 @@ Mevcut model listesi şunları içerir:
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -90,10 +88,8 @@ Aşağıdaki tablo, tipik Go kullanım modellerine dayalı tahmini bir istek say
 | GLM-5              | 1,150              | 2,880          | 5,750       |
 | Kimi K2.5          | 1,850              | 4,630          | 9,250       |
 | Kimi K2.6          | 1,150              | 2,880          | 5,750       |
-| MiMo-V2-Pro        | 1,290              | 3,225          | 6,450       |
-| MiMo-V2-Omni       | 2,150              | 5,450          | 10,900      |
-| MiMo-V2.5-Pro      | 1,290              | 3,225          | 6,450       |
 | MiMo-V2.5 (≤ 256K) | 2,150              | 5,450          | 10,900      |
+| MiMo-V2.5-Pro      | 1,290              | 3,225          | 6,450       |
 | MiniMax M2.7       | 3,400              | 8,500          | 17,000      |
 | MiniMax M2.5       | 6,300              | 15,900         | 31,800      |
 | Qwen3.6 Plus       | 3,300              | 8,200          | 16,300      |
@@ -110,10 +106,8 @@ Tahminler, gözlemlenen ortalama istek modellerine dayanmaktadır:
 - MiniMax M2.7/M2.5 — İstek başına 300 girdi, 55.000 önbelleğe alınmış, 125 çıktı token'ı
 - Qwen3.5 Plus — İstek başına 410 girdi, 47.000 önbelleğe alınmış, 140 çıktı token'ı
 - Qwen3.6 Plus — İstek başına 500 girdi, 57.000 önbelleğe alınmış, 190 çıktı token'ı
-- MiMo-V2-Pro — İstek başına 350 girdi, 41.000 önbelleğe alınmış, 250 çıktı token'ı
-- MiMo-V2-Omni — İstek başına 1000 girdi, 60.000 önbelleğe alınmış, 140 çıktı token'ı
-- MiMo-V2.5-Pro — İstek başına 350 girdi, 41.000 önbelleğe alınmış, 250 çıktı token'ı
 - MiMo-V2.5 — İstek başına 1000 girdi, 60.000 önbelleğe alınmış, 140 çıktı token'ı
+- MiMo-V2.5-Pro — İstek başına 350 girdi, 41.000 önbelleğe alınmış, 250 çıktı token'ı
 
 Mevcut kullanımınızı **<a href={console}>konsoldan</a>** takip edebilirsiniz.
 
@@ -143,10 +137,8 @@ Go modellerine aşağıdaki API uç noktaları aracılığıyla da erişebilirsi
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/tr/zen.mdx b/packages/web/src/content/docs/tr/zen.mdx
index 3e53ba40d4..36c1bfc66e 100644
--- a/packages/web/src/content/docs/tr/zen.mdx
+++ b/packages/web/src/content/docs/tr/zen.mdx
@@ -161,7 +161,7 @@ Kullandıkça öde modelini destekliyoruz. Aşağıda **1M token başına** fiya
 | GPT 5.1 Codex Mini                | $0.25  | $2.00   | $0.025      | -            |
 | GPT 5                             | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
-| GPT 5 Nano                        | Free   | Free    | Free        | -            |
+| GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |
 
 Kullanım geçmişinizde _Claude Haiku 3.5_ görebilirsiniz. Bu, oturum başlıklarınızı oluşturmak için kullanılan [düşük maliyetli bir modeldir](/docs/config/#models).
 
diff --git a/packages/web/src/content/docs/zen.mdx b/packages/web/src/content/docs/zen.mdx
index 58baceb258..333e74434b 100644
--- a/packages/web/src/content/docs/zen.mdx
+++ b/packages/web/src/content/docs/zen.mdx
@@ -172,7 +172,7 @@ We support a pay-as-you-go model. Below are the prices **per 1M tokens**.
 | GPT 5.1 Codex Mini                | $0.25  | $2.00   | $0.025      | -            |
 | GPT 5                             | $1.07  | $8.50   | $0.107      | -            |
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107      | -            |
-| GPT 5 Nano                        | Free   | Free    | Free        | -            |
+| GPT 5 Nano                        | $0.05  | $0.40   | $0.005      | -            |
 
 You might notice _Claude Haiku 3.5_ in your usage history. This is a [low cost model](/docs/config/#models) that's used to generate the titles of your sessions.
 
diff --git a/packages/web/src/content/docs/zh-cn/go.mdx b/packages/web/src/content/docs/zh-cn/go.mdx
index 8bd90d5fbf..17934ee2a0 100644
--- a/packages/web/src/content/docs/zh-cn/go.mdx
+++ b/packages/web/src/content/docs/zh-cn/go.mdx
@@ -57,10 +57,8 @@ OpenCode Go 的工作方式与 OpenCode 中的其他提供商一样。
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -90,10 +88,8 @@ OpenCode Go 包含以下限制：
 | GLM-5              | 1,150           | 2,880      | 5,750      |
 | Kimi K2.6          | 1,150           | 2,880      | 5,750      |
 | Kimi K2.5          | 1,850           | 4,630      | 9,250      |
-| MiMo-V2-Pro        | 1,290           | 3,225      | 6,450      |
-| MiMo-V2-Omni       | 2,150           | 5,450      | 10,900     |
-| MiMo-V2.5-Pro      | 1,290           | 3,225      | 6,450      |
 | MiMo-V2.5 (≤ 256K) | 2,150           | 5,450      | 10,900     |
+| MiMo-V2.5-Pro      | 1,290           | 3,225      | 6,450      |
 | Qwen3.6 Plus       | 3,300           | 8,200      | 16,300     |
 | MiniMax M2.7       | 3,400           | 8,500      | 17,000     |
 | MiniMax M2.5       | 6,300           | 15,900     | 31,800     |
@@ -107,10 +103,8 @@ OpenCode Go 包含以下限制：
 - Kimi K2.5/K2.6 — 每次请求 870 个输入 token，55,000 个缓存 token，200 个输出 token
 - DeepSeek V4 Pro — 每次请求 750 个输入 token，82,000 个缓存 token，290 个输出 token
 - DeepSeek V4 Flash — 每次请求 790 个输入 token，68,000 个缓存 token，280 个输出 token
-- MiMo-V2-Pro — 每次请求 350 个输入 token，41,000 个缓存 token，250 个输出 token
-- MiMo-V2-Omni — 每次请求 1000 个输入 token，60,000 个缓存 token，140 个输出 token
-- MiMo-V2.5-Pro — 每次请求 350 个输入 token，41,000 个缓存 token，250 个输出 token
 - MiMo-V2.5 — 每次请求 1000 个输入 token，60,000 个缓存 token，140 个输出 token
+- MiMo-V2.5-Pro — 每次请求 350 个输入 token，41,000 个缓存 token，250 个输出 token
 - MiniMax M2.7/M2.5 — 每次请求 300 个输入 token，55,000 个缓存 token，125 个输出 token
 - Qwen3.5 Plus — 每次请求 410 个输入 token，47,000 个缓存 token，140 个输出 token
 - Qwen3.6 Plus — 每次请求 500 个输入 token，57,000 个缓存 token，190 个输出 token
@@ -143,10 +137,8 @@ OpenCode Go 包含以下限制：
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/zh-cn/zen.mdx b/packages/web/src/content/docs/zh-cn/zen.mdx
index 03124a34f4..9ad7e6b53d 100644
--- a/packages/web/src/content/docs/zh-cn/zen.mdx
+++ b/packages/web/src/content/docs/zh-cn/zen.mdx
@@ -161,7 +161,7 @@ https://opencode.ai/zen/v1/models
 | GPT 5.1 Codex Mini                | $0.25  | $2.00   | $0.025   | -        |
 | GPT 5                             | $1.07  | $8.50   | $0.107   | -        |
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107   | -        |
-| GPT 5 Nano                        | Free   | Free    | Free     | -        |
+| GPT 5 Nano                        | $0.05  | $0.40   | $0.005   | -        |
 
 你可能会在使用记录中看到 _Claude Haiku 3.5_。这是一个[低成本模型](/docs/config/#models)，用于生成会话标题。
 
diff --git a/packages/web/src/content/docs/zh-tw/go.mdx b/packages/web/src/content/docs/zh-tw/go.mdx
index 3bf4618bc5..c4589716f2 100644
--- a/packages/web/src/content/docs/zh-tw/go.mdx
+++ b/packages/web/src/content/docs/zh-tw/go.mdx
@@ -57,10 +57,8 @@ OpenCode Go 的運作方式與 OpenCode 中的任何其他供應商相同。
 - **GLM-5.1**
 - **Kimi K2.5**
 - **Kimi K2.6**
-- **MiMo-V2-Pro**
-- **MiMo-V2-Omni**
-- **MiMo-V2.5-Pro**
 - **MiMo-V2.5**
+- **MiMo-V2.5-Pro**
 - **MiniMax M2.5**
 - **Qwen3.5 Plus**
 - **Qwen3.6 Plus**
@@ -90,10 +88,8 @@ OpenCode Go 包含以下限制：
 | GLM-5              | 1,150           | 2,880      | 5,750      |
 | Kimi K2.5          | 1,850           | 4,630      | 9,250      |
 | Kimi K2.6          | 1,150           | 2,880      | 5,750      |
-| MiMo-V2-Pro        | 1,290           | 3,225      | 6,450      |
-| MiMo-V2-Omni       | 2,150           | 5,450      | 10,900     |
-| MiMo-V2.5-Pro      | 1,290           | 3,225      | 6,450      |
 | MiMo-V2.5 (≤ 256K) | 2,150           | 5,450      | 10,900     |
+| MiMo-V2.5-Pro      | 1,290           | 3,225      | 6,450      |
 | Qwen3.6 Plus       | 3,300           | 8,200      | 16,300     |
 | MiniMax M2.7       | 3,400           | 8,500      | 17,000     |
 | MiniMax M2.5       | 6,300           | 15,900     | 31,800     |
@@ -110,10 +106,8 @@ OpenCode Go 包含以下限制：
 - MiniMax M2.7/M2.5 — 每次請求 300 個輸入 token、55,000 個快取 token、125 個輸出 token
 - Qwen3.5 Plus — 每次請求 410 個輸入 token、47,000 個快取 token、140 個輸出 token
 - Qwen3.6 Plus — 每次請求 500 個輸入 token、57,000 個快取 token、190 個輸出 token
-- MiMo-V2-Pro — 每次請求 350 個輸入 token、41,000 個快取 token、250 個輸出 token
-- MiMo-V2-Omni — 每次請求 1000 個輸入 token、60,000 個快取 token、140 個輸出 token
-- MiMo-V2.5-Pro — 每次請求 350 個輸入 token、41,000 個快取 token、250 個輸出 token
 - MiMo-V2.5 — 每次請求 1000 個輸入 token、60,000 個快取 token、140 個輸出 token
+- MiMo-V2.5-Pro — 每次請求 350 個輸入 token、41,000 個快取 token、250 個輸出 token
 
 您可以在 **<a href={console}>console</a>** 中追蹤您目前的使用量。
 
@@ -143,10 +137,8 @@ OpenCode Go 包含以下限制：
 | Kimi K2.6         | kimi-k2.6         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Pro   | deepseek-v4-pro   | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Pro       | mimo-v2-pro       | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2-Omni      | mimo-v2-omni      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
-| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiMo-V2.5         | mimo-v2.5         | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
+| MiMo-V2.5-Pro     | mimo-v2.5-pro     | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
 | MiniMax M2.7      | minimax-m2.7      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | MiniMax M2.5      | minimax-m2.5      | `https://opencode.ai/zen/go/v1/messages`         | `@ai-sdk/anthropic`         |
 | Qwen3.6 Plus      | qwen3.6-plus      | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/alibaba`           |
diff --git a/packages/web/src/content/docs/zh-tw/zen.mdx b/packages/web/src/content/docs/zh-tw/zen.mdx
index ebd48dea8f..9511bd9e24 100644
--- a/packages/web/src/content/docs/zh-tw/zen.mdx
+++ b/packages/web/src/content/docs/zh-tw/zen.mdx
@@ -166,7 +166,7 @@ https://opencode.ai/zen/v1/models
 | GPT 5.1 Codex Mini                | $0.25  | $2.00   | $0.025   | -        |
 | GPT 5                             | $1.07  | $8.50   | $0.107   | -        |
 | GPT 5 Codex                       | $1.07  | $8.50   | $0.107   | -        |
-| GPT 5 Nano                        | Free   | Free    | Free     | -        |
+| GPT 5 Nano                        | $0.05  | $0.40   | $0.005   | -        |
 
 你可能會在使用紀錄中看到 _Claude Haiku 3.5_。這是一個[低成本模型](/docs/config/#models)，
 會用來產生工作階段的標題。
diff --git a/sdks/vscode/package.json b/sdks/vscode/package.json
index c78e2a1486..3eaca42fb7 100644
--- a/sdks/vscode/package.json
+++ b/sdks/vscode/package.json
@@ -2,7 +2,7 @@
   "name": "opencode",
   "displayName": "opencode",
   "description": "opencode for VS Code",
-  "version": "1.14.39",
+  "version": "1.14.41",
   "publisher": "sst-dev",
   "repository": {
     "type": "git",
diff --git a/sst-env.d.ts b/sst-env.d.ts
index bb6287a157..e75c54d056 100644
--- a/sst-env.d.ts
+++ b/sst-env.d.ts
@@ -50,6 +50,10 @@ declare module "sst" {
       "type": "sst.cloudflare.SolidStart"
       "url": string
     }
+    "DISCORD_INCIDENT_WEBHOOK_URL": {
+      "type": "sst.sst.Secret"
+      "value": string
+    }
     "DISCORD_SUPPORT_BOT_TOKEN": {
       "type": "sst.sst.Secret"
       "value": string
@@ -110,6 +114,10 @@ declare module "sst" {
       "type": "sst.sst.Secret"
       "value": string
     }
+    "HoneycombWebhookSecret": {
+      "type": "random.index/randomPassword.RandomPassword"
+      "value": string
+    }
     "LogProcessor": {
       "type": "sst.cloudflare.Worker"
     }
diff --git a/sst.config.ts b/sst.config.ts
index b8e56473bc..696a6fa768 100644
--- a/sst.config.ts
+++ b/sst.config.ts
@@ -11,7 +11,9 @@ export default $config({
         stripe: {
           apiKey: process.env.STRIPE_SECRET_KEY!,
         },
+        random: "4.19.2",
         planetscale: "0.4.1",
+        honeycomb: "0.49.0",
       },
     }
   },
@@ -19,5 +21,8 @@ export default $config({
     await import("./infra/app.js")
     await import("./infra/console.js")
     await import("./infra/enterprise.js")
+    if ($app.stage === "production" || $app.stage === "vimtor") {
+      await import("./infra/monitoring.js")
+    }
   },
 })