Skip to content

feat: upgrade default MiniMax model to M3#1512

Open
Hansen1018 wants to merge 54 commits into
NoFxAiOS:devfrom
Hansen1018:feat/minimax-m3
Open

feat: upgrade default MiniMax model to M3#1512
Hansen1018 wants to merge 54 commits into
NoFxAiOS:devfrom
Hansen1018:feat/minimax-m3

Conversation

@Hansen1018
Copy link
Copy Markdown
Contributor

Summary

  • Problem: Upgrade default MiniMax model to the new M3 flagship tier
  • What changed: Updated default MiniMax model from M2.7/M2.5 to M3 across all provider configs, added MiniMax-M3 to pricing at $0.20/call
  • What did NOT change (scope boundary): All other AI models and providers remain unchanged. Older MiniMax model versions still available via custom model field.

M3 Model Specifications (per official docs)

  • Architecture: MiniMax Sparse Attention (MSA)
  • Context window: up to 1M tokens (at least 512K guaranteed usable)
  • Modalities: Native multimodal — image and video input, can operate computer desktop
  • Coding: Exceeds GPT-5.5 on SWE-Bench Pro
  • Agent (BrowseComp): 83.5 score (vs Opus 4.7: 79.3)

Reference: https://minimaxi.com/M3 (released 2026-06-01)

Change Type

  • Feature

Scope

  • MCP / AI clients
  • API / server
  • Web UI / frontend

Testing

  • go build ./... passes
  • npm run build passes (web)
  • go test ./... passes

Compatibility

  • Backward compatible? (Yes - older MiniMax model versions still selectable via custom model field)
  • Migration needed? (No)
  • Config/env changes? (No)

Hansen1018 and others added 24 commits March 27, 2026 21:08
- mcp/provider/minimax.go: DefaultMiniMaxModel = MiniMax-M3
- mcp/providers.go: DefaultMiniMaxModel = MiniMax-M3
- agent/agent.go: minimax provider default model = MiniMax-M3
- agent/model_provider_catalog.go: MiniMax DefaultModel = MiniMax-M3
- api/handler_ai_model.go: minimax defaultModel = MiniMax-M3
- web/src/components/trader/model-constants.ts: minimax defaultModel = MiniMax-M3
- store/ai_charge.go: MiniMax-M3 price at $0.20/call

Reference: https://minimaxi.com/M3 (released 2026-06-01)
- MiniMax Sparse Attention (MSA) architecture
- 1M tokens context window (512K guaranteed usable)
- Native multimodal: image and video input
- Coding capability exceeds GPT-5.5
- BrowseComp agent score: 83.5 (vs Opus 4.7: 79.3)
@github-actions
Copy link
Copy Markdown

github-actions Bot commented Jun 7, 2026

🤖 Advisory Check Results

These are advisory checks to help improve code quality. They won't block your PR from being merged.

📋 PR Information

Title Format: ✅ Good - Follows Conventional Commits
PR Size: 🟢 Small (13 lines: +7 -6)

🔧 Backend Checks

Go Formatting: ⚠️ Needs formatting

Files needing formatting
agent/execution_state.go
agent/model_provider_catalog.go
agent/skill_dag_runtime.go
agent/stock.go
api/errors.go
api/handler_wallet.go
cmd/e2e_builder_fee/main.go
kernel/formatter.go
kernel/grid_engine.go
kernel/validate_test.go

Go Vet: ✅ Good
Tests: ✅ Passed

Fix locally:

go fmt ./...      # Format code
go vet ./...      # Check for issues
go test ./...     # Run tests

⚛️ Frontend Checks

Build & Type Check: ✅ Success

Fix locally:

cd web
npm run build  # Test build (includes type checking)

📖 Resources

Questions? Feel free to ask in the comments! 🙏


These checks are advisory and won't block your PR from being merged. This comment is automatically generated from pr-checks-run.yml.

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