Add MiniMax as LLM provider with examples, docs, and tests

[octo-patch]

Summary

• Add MiniMax as an OpenAI-compatible LLM provider using the existing OpenAIModel + OpenAIProvider pattern
• Support MiniMax-M2.7 (peak performance) and MiniMax-M2.7-highspeed (faster, more agile) models with 204K context windows
• Add provider-specific examples following the aimlapi pattern: basic agent, structured output, and tool usage
• Update LLM configuration docs with MiniMax usage guide
• Add unit tests for model configuration and integration tests for API calls

Changes

File	Description
examples/provider_specific/minimax/run_agent.py	Basic agent example using MiniMax
examples/provider_specific/minimax/structured_output.py	Structured output extraction example
examples/provider_specific/minimax/tools_agent.py	Agent with tool calling example
docs/guides/configure-llms.mdx	Added MiniMax to provider list and usage guide
tests/basic/actors/test_minimax_provider.py	Unit tests (6) + integration tests (3)

Test plan

• Unit tests pass: model configuration, provider base URL, temperature settings, tool preservation
• Integration tests pass: simple run, structured output, highspeed model (all verified against live MiniMax API)
• Examples run successfully with MINIMAX_API_KEY set