feat: add MiniMax as a native LLM and embedding provider (M2.7)

[octo-patch]

Summary

• Add MiniMax as a native LLM provider using their OpenAI-compatible API (https://api.minimax.io/v1)
• Add MiniMax embedding support (e.g. embo-01) via the same OpenAI-compatible endpoint
• Recommend MiniMax-M2.7 and M2.7-highspeed as the default models (upgraded from M2.5)
• Add documentation with configuration examples for all MiniMax models

Details

MiniMax offers powerful large language models with an OpenAI-compatible API. This PR adds first-class support so users can configure MiniMax models directly:

MINIMAX_API_KEY=[Your Key]
FAST_LLM=minimax:MiniMax-M2.7-highspeed
SMART_LLM=minimax:MiniMax-M2.7
STRATEGIC_LLM=minimax:MiniMax-M2.7
EMBEDDING=minimax:embo-01

Available Models

• MiniMax-M2.7 — latest flagship model with improved reasoning
• MiniMax-M2.7-highspeed — optimized for speed
• MiniMax-M2.5 — 204K context, general-purpose model
• MiniMax-M2.5-highspeed — 204K context, optimized for speed

Implementation

• LLM: Uses langchain_openai.ChatOpenAI with openai_api_base=https://api.minimax.io/v1
• Embeddings: Uses langchain_openai.OpenAIEmbeddings with the same base URL
• No new dependencies required — reuses existing langchain-openai package

Test Plan

• Verified MiniMax-M2.7 model responds correctly via API
• Verified MiniMax-M2.7-highspeed model responds correctly via API
• Verified embedding endpoint works with embo-01 model
• Documentation updated with configuration examples

[octo-patch]

Thanks for the interest! Let me know if you have any questions about the MiniMax integration.

[wisdomdaoleo]

🔬 GPT Researcher 很棒！AI 研究自动化非常有价值。

如果你对各种 AI Agent 感兴趣，欢迎加入讨论：

🤖 Bot: @xinhaovip2_bot
👥 群组: https://t.me/atoxinhaovip2025