Skip to content

[feat] [model] add MiniMax as LLM provider protocol (M3 default)#466

Open
octo-patch wants to merge 1 commit into
coze-dev:mainfrom
octo-patch:feature/add-minimax-provider
Open

[feat] [model] add MiniMax as LLM provider protocol (M3 default)#466
octo-patch wants to merge 1 commit into
coze-dev:mainfrom
octo-patch:feature/add-minimax-provider

Conversation

@octo-patch

@octo-patch octo-patch commented Mar 21, 2026

Copy link
Copy Markdown

Summary

  • Add MiniMax as the 10th LLM provider protocol in coze-loop model system
  • MiniMax provides OpenAI-compatible API, integrated via existing eino-ext openai component with default base URL https://api.minimax.io/v1
  • Supported models: MiniMax-M3 (default, 512K context / 128K output, image input), MiniMax-M2.7, MiniMax-M2.7-highspeed

Changes

Backend (Go)

  • entity/manage.go: Add ProtocolMiniMax constant and ProtocolConfigMiniMax struct
  • eino/init.go: Add miniMaxBuilder function using OpenAI-compatible eino-ext component
  • kitex_gen manage.go: Add ProtocolMinimax constant

IDL

  • manage.thrift: Add protocol_minimax and ProtocolConfigMiniMax struct

Frontend (TypeScript)

  • manage.ts: Add protocol_minimax enum value and ProtocolConfigMiniMax interface

Test plan

  • 10 unit tests for minimax builder (various configs, error cases, model variants — default model MiniMax-M3)
  • 1 factory integration test (eino_minimax case in TestFactoryImpl_CreateLLM)
  • 3 integration tests with real MiniMax API (M3 generate, M2.7-highspeed generate, M3 stream)
  • All existing tests pass without regression

mocayo
mocayo previously approved these changes Mar 23, 2026
@mocayo mocayo changed the title feat(llm): add MiniMax as LLM provider protocol [feat] [model] add MiniMax as LLM provider protocol Mar 23, 2026
Add MiniMax as the 10th LLM provider in coze-loop model protocol system.
MiniMax offers OpenAI-compatible API and is integrated via the existing
eino-ext openai component with MiniMax default base URL.

Changes:
- Add ProtocolMiniMax constant and ProtocolConfigMiniMax struct in entity
- Add miniMaxBuilder in eino/init.go using openai component with MiniMax defaults
- Update thrift IDL with protocol_minimax and ProtocolConfigMiniMax
- Update kitex_gen constants and frontend TypeScript types
- Add 10 unit tests and 3 integration tests

Supported models: MiniMax-M3 (default, 512K context / 128K output),
MiniMax-M2.7, MiniMax-M2.7-highspeed
@octo-patch octo-patch force-pushed the feature/add-minimax-provider branch from 8a3941e to a0ff4d8 Compare June 5, 2026 07:18
@octo-patch octo-patch changed the title [feat] [model] add MiniMax as LLM provider protocol [feat] [model] add MiniMax as LLM provider protocol (M3 default) Jun 5, 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.

2 participants