fix(ai-gateway): default MiniMax M3 to thinking

[lambertjosh]

Summary

Keep the common Kilo Gateway thinking / instant labels for MiniMax M3, but return thinking first so new IDE selections default to adaptive thinking. instant remains available as the explicit disabled-thinking option, and older MiniMax models retain their existing instant-first ordering.

Verification

• No manual UI verification; this is a model-catalog metadata change covered by focused gateway tests.

Visual Changes

N/A

Reviewer Notes

This applies to ordinary Kilo Gateway minimax/minimax-m3 routing, including Token Plan Plus-installed credentials, and direct OpenCode Go M3 models. The companion Kilo CLI change in Kilo-Org/kilocode#11562 ensures an implicit/default Gateway request also uses adaptive thinking; selecting instant explicitly overrides it with disabled thinking.

[kilo-code-bot]

Code Review Summary

Status: No Issues Found | Recommendation: Merge

Files Reviewed (2 files)

• apps/web/src/lib/ai-gateway/providers/model-settings.ts
• apps/web/src/lib/ai-gateway/providers/model-settings.test.ts

Previous Review Summaries (2 snapshots, latest commit 7fdc188)

Current summary above is authoritative. Previous snapshots are kept for context only.

Previous review (commit 7fdc188)

Status: No Issues Found | Recommendation: Merge

Files Reviewed (2 files)

• apps/web/src/lib/ai-gateway/providers/model-settings.ts
• apps/web/src/lib/ai-gateway/providers/model-settings.test.ts

Previous review (commit cc51c2d)

Status: No Issues Found | Recommendation: Merge

Files Reviewed (2 files)

• apps/web/src/lib/ai-gateway/providers/model-settings.ts
• apps/web/src/lib/ai-gateway/providers/model-settings.test.ts

---

_{Reviewed by gpt-5.4-20260305 · Input: 95.4K · Output: 7.9K · Cached: 665.1K}

_{Review guidance: REVIEW.md from base branch main}