feat: add MiniMax as a supported LLM provider

[octo-patch]

Summary

• Add MiniMax as a new LLM provider with OpenAI-compatible API integration
• Register MiniMax models (MiniMax-M1, M2, M2.5, M2.5-highspeed) in the model token registry
• Add provider routing in the abstract graph factory method
• Update README to list MiniMax as a supported provider

Details

MiniMax offers an OpenAI-compatible API, so the integration follows the same pattern as DeepSeek and xAI — a lightweight wrapper around ChatOpenAI that sets the correct API base URL.

Usage

graph_config = {
    "llm": {
        "api_key": "YOUR_MINIMAX_API_KEY",
        "model": "minimax/MiniMax-M2.5",
    },
    "verbose": True,
    "headless": False,
}

Models supported

Model	Context Window
MiniMax-M1	1,000,000
MiniMax-M1-40k	40,000
MiniMax-M2	204,000
MiniMax-M2.5	204,000
MiniMax-M2.5-highspeed	204,000

Test plan

• MiniMax model class imports correctly
• Model tokens registered and retrievable
• API base URL set correctly to https://api.minimax.io/v1
• Existing tests pass (16/16 that are not pre-existing failures)

[github-actions]

🎉 This PR is included in version 1.74.0 🎉

The release is available on:

• v1.74.0
• GitHub release

Your semantic-release bot 📦🚀