Skip to content

feat: add MiniMax as LLM provider with M3 model support (M3 default, M2.7 kept)#2639

Open
octo-patch wants to merge 3 commits into
coze-dev:mainfrom
octo-patch:feature/add-minimax-provider
Open

feat: add MiniMax as LLM provider with M3 model support (M3 default, M2.7 kept)#2639
octo-patch wants to merge 3 commits into
coze-dev:mainfrom
octo-patch:feature/add-minimax-provider

Conversation

@octo-patch
Copy link
Copy Markdown

@octo-patch octo-patch commented Mar 17, 2026

Summary

  • Add MiniMax as a first-class LLM provider via OpenAI-compatible eino-ext integration
  • Support MiniMax-M3 (default), MiniMax-M2.7, and MiniMax-M2.7-highspeed models
  • M3 is the latest model with 512K context, 128K max output, and image input support
  • Temperature clamping (MiniMax requires temperature > 0)
  • Response format forced to text (MiniMax does not support JSON mode)
  • Full model builder with unit and integration tests

Changed Files

  • backend/bizpkg/llm/modelbuilder/minimax.go - MiniMax model builder
  • backend/bizpkg/llm/modelbuilder/minimax_test.go - Unit + integration tests
  • backend/bizpkg/llm/modelbuilder/model_builder.go - Registration
  • backend/conf/model/model_meta.json - Model metadata with M3/M2.7 entries
  • backend/conf/model/template/model_template_minimax.yaml - Template config (M3 default)
  • helm/charts/opencoze/files/conf/model/model_meta.json - Helm chart model metadata
  • Multiple auto-generated IDL files with ModelClass_MiniMax enum

Model Configuration

Model Context Max Output Image Input Default
MiniMax-M3 512K 128K yes yes
MiniMax-M2.7 204K 16K no no
MiniMax-M2.7-highspeed 204K 16K no no

Older models (M2.5, M2.5-highspeed) have been removed.

Test Plan

  • Unit tests for temperature clamping
  • Unit tests for model builder config
  • Unit tests for model builder registration
  • Integration tests for M3, M2.7, M2.7-highspeed models
  • JSON and YAML file validation

Add MiniMax (MiniMax-M2.5 and MiniMax-M2.5-highspeed) as a first-class
LLM provider using the OpenAI-compatible API endpoint.

Changes:
- Add minimax.go model builder with temperature clamping (0,1.0]
- Register MiniMax in model builder factory map
- Add ProtocolMiniMax protocol constant and mapping
- Add MiniMax to provider list with icon and description
- Add model_template_minimax.yaml with 204K context config
- Add MiniMax models to model_meta.json (default, M2.5, M2.5-highspeed)
- Add comprehensive unit tests (7 test cases) and integration tests
- Update README, README.zh_CN, and CLAUDE.md with MiniMax in provider list
Add MiniMax-M2.7 and MiniMax-M2.7-highspeed model entries to model
metadata and update the default template model to M2.7 with advanced
reasoning capabilities and 16K output tokens.
@CLAassistant
Copy link
Copy Markdown

CLA assistant check
Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
1 out of 2 committers have signed the CLA.

✅ octo-patch
❌ PR Bot


PR Bot seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
You have signed the CLA already but the status is still pending? Let us recheck it.

@octo-patch octo-patch changed the title feat: add MiniMax as LLM provider feat: add MiniMax as LLM provider with M2.7 model support Mar 18, 2026
- Add MiniMax-M3 to model list and set as default (512K context, 128K max output, image input)
- Keep MiniMax-M2.7 and MiniMax-M2.7-highspeed as alternatives
- Remove older models (M2.5/M2.5-highspeed)
- Update related unit tests and integration test model list
- Sync helm chart model_meta.json with backend config
@octo-patch octo-patch changed the title feat: add MiniMax as LLM provider with M2.7 model support feat: add MiniMax as LLM provider with M3 model support (M3 default, M2.7 kept) Jun 2, 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