Skip to content

Add MiniMax provider preset with global and China endpoints to LLMClient - #2

Open
octo-patch wants to merge 2 commits into
AtomFlow-AI:mainfrom
octo-patch:octo/20260731-provider-add-recvqd3y1gYExC
Open

Add MiniMax provider preset with global and China endpoints to LLMClient#2
octo-patch wants to merge 2 commits into
AtomFlow-AI:mainfrom
octo-patch:octo/20260731-provider-add-recvqd3y1gYExC

Conversation

@octo-patch

Copy link
Copy Markdown

Reason: Add MiniMax as a native provider in LLMClient with global and China endpoint presets and MiniMax-M3 / MiniMax-M2.7 model support.

Changes

  • Add to mapping the MiniMax provider to its documented OpenAI Chat Completions base URLs per region ( -> , -> ).
  • Add and kwargs (and matching / env vars) to , with resolving the base URL from an explicit value, then env, then a provider preset, then the OpenAI default. An explicit keeps winning, and the existing image-capable chat request path is unchanged.
  • Add covering both regional endpoints, both MiniMax-M3 and MiniMax-M2.7, case-insensitive provider lookup, explicit-URL override, region fallback, env-based configuration, the assembled URL, and the unchanged image helper.

Verification

  • Ran (11 passed).

@octo-patch

Copy link
Copy Markdown
Author

Updated MiniMax provider support with Anthropic Messages endpoints for both regions, current model defaults, and image-content conversion. I ran Python 3.9/3.14 compilation, git diff checks, and offline request/response smoke tests; the focused pytest suite was not run because this worker lacks pytest and rdkit.

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