feat: add MiniMax as LLM provider (M3 default)

[octo-patch]

Summary

• Add MiniMax AI as a first-class LLM provider (MiniMax-M3 default, MiniMax-M2.7, MiniMax-M2.7-highspeed) via OpenAI-compatible API at api.minimax.io/v1
• Add MINIMAX elif branch in get_llm() factory using ChatOpenAI with custom base_url
• Add MiniMax model config examples in backend/example.env and default model entries in frontend/src/utils/Constants.ts
• Update README to list MiniMax as a supported LLM

Changes

File	Change
backend/src/llm.py	Add MINIMAX elif branch in get_llm() factory
backend/example.env	Add MiniMax M3 / M2.7 / M2.7-highspeed config examples
frontend/src/utils/Constants.ts	Add MiniMax models to default LLM list (M3 default)
README.md	Add MiniMax to supported LLMs list
backend/tests/test_minimax_provider.py	Unit tests for routing, config, params (M3 + M2.7)
backend/tests/test_minimax_integration.py	Integration tests (API connectivity, M3 + M2.7)

Configuration

# backend/.env
LLM_MODEL_CONFIG_minimax_m3="MiniMax-M3,your_minimax_api_key"
LLM_MODEL_CONFIG_minimax_m2.7="MiniMax-M2.7,your_minimax_api_key"
LLM_MODEL_CONFIG_minimax_m2.7_highspeed="MiniMax-M2.7-highspeed,your_minimax_api_key"

Why MiniMax-M3

MiniMax-M3 is the latest MiniMax model: 512K context window, up to 128K output, image input support. Set as the default in the frontend selector; M2.7 and M2.7-highspeed remain available as alternatives.

Test plan

• Unit tests pass (model routing, env config, source code validation) — M3 + M2.7 coverage
• Integration tests pass (M3 chat, M2.7 chat, M2.7-highspeed chat, structured output support)
• Manual verification: add MiniMax config to .env, select minimax_m3 in frontend, run graph extraction