Skip to content

provider: add MiniMax as LLM provider (rebased & signed)#1650

Open
kapelame wants to merge 1 commit into
Portkey-AI:mainfrom
kapelame:feature/add-minimax-provider
Open

provider: add MiniMax as LLM provider (rebased & signed)#1650
kapelame wants to merge 1 commit into
Portkey-AI:mainfrom
kapelame:feature/add-minimax-provider

Conversation

@kapelame
Copy link
Copy Markdown

Reopens #1571 with rebase + DCO sign-off.

  • Same code as feat: add MiniMax as LLM provider #1571 (already approved by @narengogi)
  • Rebased onto current main
  • Signed-off-by: trailer added (DCO)
  • Co-authored-by: PR Bot <pr-bot@minimaxi.com> to preserve credit to the original author (@octo-patch)

Coordinated with @narengogi via DM. Once this is merged, #1571 can be closed.

cc @narengogi @octo-patch

Add MiniMax as a first-class LLM provider to the Portkey AI Gateway,
supporting MiniMax-M2.7 and MiniMax-M2.7-highspeed models via their
OpenAI-compatible API at https://api.minimax.io/v1.

- Add provider files (api.ts, chatComplete.ts, index.ts) with
  streaming support and response transforms
- Register provider in globals, provider index, providers.json,
  and models.json
- Use open-ai-base for chat completion params with temperature
  default of 1.0
- Exclude unsupported params: logit_bias, logprobs, top_logprobs
- Add 22 unit tests covering config, transforms, and streaming
- Add 3 integration tests with real API validation
- Add test variables entry for MINIMAX_API_KEY

Co-authored-by: PR Bot <pr-bot@minimaxi.com>
Signed-off-by: kapelame <kapelame@users.noreply.github.com>
@kapelame kapelame force-pushed the feature/add-minimax-provider branch from ccc2ebd to 1a568bb Compare May 14, 2026 02:32
@kapelame
Copy link
Copy Markdown
Author

@narengogi commit is now SSH-signed (GitHub verifies as ✅). Same code as #1571 (your approval there). Ready for re-review whenever you've got a moment 🙏

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.

2 participants