Skip to content

feat: add MiniMax Flash and M3 pricing fallbacks - #81

Open
octo-patch wants to merge 2 commits into
xaspx:mainfrom
octo-patch:octo/20260907-model-add-recvthsb3vJWay
Open

octo-patch wants to merge 2 commits into
xaspx:mainfrom
octo-patch:octo/20260907-model-add-recvthsb3vJWay

Conversation

@octo-patch

@octo-patch octo-patch commented Sep 7, 2026 •

Copy link
Copy Markdown

Reason: Add MiniMax-M3.1-Flash and MiniMax-M3 custom pricing fallbacks for accurate usage costs.

The pricing helper now covers MiniMax-M3.1-Flash at the documented direct rates ($0.15 input, $0.60 output, $0.015 cache-read per million tokens) and MiniMax-M3 at the documented standard baseline ($0.30/$1.20/$0.06). It preserves existing saved/custom model behavior, free-model handling, provider-aware SDK lookup, and legacy pricing entries. The M3 long-context threshold is retained as a documented limitation because the published value does not define its exact integer expansion.

The pricing logic is isolated in lib/pricing.js so the SDK boundary can be tested offline with a mocked @pydantic/genai-prices module. This repository owns usage-cost calculation; model catalogs/defaults and MiniMax HTTP transport remain delegated to Hermes Agent.

Checks:

  • npm test (16 passing unit tests)
  • npm run build
  • node --check server.js
  • node --check lib/pricing.js
  • node --check test/pricing.test.js
  • git diff --check

@octo-patch octo-patch changed the title Add MiniMax M3 pricing fallback feat: add MiniMax Flash and M3 pricing fallbacks Oct 21, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant