Skip to content

feat: upgrade MiniMax default model to M3 (keep M2.7, drop M2.5)#2362

Open
octo-patch wants to merge 3 commits into
songquanpeng:mainfrom
octo-patch:feature/add-minimax-m25-models
Open

feat: upgrade MiniMax default model to M3 (keep M2.7, drop M2.5)#2362
octo-patch wants to merge 3 commits into
songquanpeng:mainfrom
octo-patch:feature/add-minimax-m25-models

Conversation

@octo-patch
Copy link
Copy Markdown

@octo-patch octo-patch commented Mar 15, 2026

Summary

Upgrade MiniMax model configuration to use M3 as the default model.

Changes

  • Add MiniMax-M3 to the model selection list and set as default (first entry)
  • Keep MiniMax-M2.7 and MiniMax-M2.7-highspeed as alternatives
  • Remove older MiniMax-M2.5 and MiniMax-M2.5-highspeed
  • Update pricing ratios for the M3 model
  • Update Berry UI channel config to expose MiniMax-M3

Model Pricing

Model Input Output
MiniMax-M3 (default) $0.6 / M tokens $2.4 / M tokens
MiniMax-M2.7 $0.3 / M tokens $1.2 / M tokens
MiniMax-M2.7-highspeed $0.6 / M tokens $2.4 / M tokens

Why

MiniMax-M3 is the latest MiniMax model, featuring a 512K context window, up to 128K max output tokens, and image input support on both the OpenAI-compatible and Anthropic-compatible endpoints.

Test plan

  • go build ./relay/adaptor/minimax/ ./relay/billing/ratio/ passes
  • go vet ./relay/adaptor/minimax/ ./relay/billing/ratio/ passes
  • M3 added to model list and is the first/default entry
  • Pricing ratios configured for M3 (ModelRatio + CompletionRatio)
  • Berry UI channel-type config exposes M3
  • M2.7 and M2.7-highspeed retained for backward compatibility

- Add MiniMax-M2.5 and MiniMax-M2.5-highspeed to model list
- Update base URL to api.minimax.io (OpenAI-compatible endpoint)
- Simplify request routing to use standard OpenAI /v1/chat/completions
- Add pricing ratios for new M2.5 models
- Update README with current MiniMax platform link
@octo-patch octo-patch changed the title feat: add MiniMax M2.5 models and update to OpenAI-compatible API feat: add MiniMax M2.7/M2.5 models and update to OpenAI-compatible API Mar 18, 2026
PR Bot and others added 2 commits March 18, 2026 11:00
- Add MiniMax-M2.7 and MiniMax-M2.7-highspeed to model list
- Add pricing ratios for M2.7 models (same as M2.5)
- Add completion ratios for M2.7 models
- Update Berry UI to show all MiniMax models including M2.7
- Keep all previous models as alternatives
- Add MiniMax-M3 to model list and set as default (first entry)
- Keep MiniMax-M2.7 and MiniMax-M2.7-highspeed
- Remove older MiniMax-M2.5 and MiniMax-M2.5-highspeed
- Update pricing ratios for M3 (input $0.6/M, output $2.4/M)
@octo-patch octo-patch changed the title feat: add MiniMax M2.7/M2.5 models and update to OpenAI-compatible API feat: upgrade MiniMax default model to M3 (keep M2.7, drop M2.5) 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.

1 participant