Skip to content

Add MiniMax provider preset with global/CN endpoints - #17

Open
octo-patch wants to merge 1 commit into
QuantaAlpha:mainfrom
octo-patch:octo/20260731-provider-add-recvqeGYWa4XAP
Open

Add MiniMax provider preset with global/CN endpoints#17
octo-patch wants to merge 1 commit into
QuantaAlpha:mainfrom
octo-patch:octo/20260731-provider-add-recvqeGYWa4XAP

Conversation

@octo-patch

Copy link
Copy Markdown

Reason: Add MiniMax as a configurable LLM provider with global and CN regional endpoints.

Changes:

  • configs/oai_config.py: register a minimax provider preset in get_api_config() (model MiniMax-M3, key MINIMAX_API_KEY, base URL defaulting to the global endpoint https://api.minimax.io/v1) and add minimax to DEFAULT_PROVIDER_PRIORITY so it participates in provider-priority fallback. The CN endpoint (https://api.minimaxi.com/v1) is documented inline and selectable via MINIMAX_BASE_URL.
  • configs/mode_config.py: add minimax to the --api-type CLI choices and to the BackendConfig.api_type docstring.
  • configs/env.example: add MINIMAX_API_KEY / MINIMAX_MODEL / MINIMAX_BASE_URL entries (defaulting to the global endpoint) and update the DEFAULT_API_PROVIDER comment to list minimax.
  • USAGE.md: document the MiniMax configuration block (global/CN endpoints) and the minimax default-provider option.

Checks performed:

  • python3 -m py_compile configs/oai_config.py configs/mode_config.py (syntax OK).
  • Import test: minimax is present in DEFAULT_PROVIDER_PRIORITY and get_api_config(), and get_llm_config(api_type='minimax') resolves model MiniMax-M3 with base URL https://api.minimax.io/v1.
  • Provider-priority fallback test (clean env, only MINIMAX_API_KEY set): validate_and_get_fallback_config() resolves to minimax.
  • create_argument_parser() exposes minimax in the --api-type choices.

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