feat: add MiniMax as first-class LLM provider

[octo-patch]

Summary

• Add MiniMax as a first-class LLM provider in the model-provider framework
• Implement mp-minimax package following the existing provider pattern (same as DeepSeek)
• Register MiniMax in provider constants, model config factory, and model params factory
• Support chat completions (unary + streaming), tool calling, vision, and thinking tags
• MiniMax-specific temperature clamping to [0, 1.0] range
• Stream requests automatically include stream_options.include_usage

Changes

• pkg/model-provider/mp-minimax/mp-minimax-llm.go - Core MiniMax LLM provider implementing ILLM interface
• pkg/model-provider/common.go - Added ProviderMiniMax constant
• pkg/model-provider/model_config_api.go - Registered MiniMax in ToModelTags(), ToModelConfig(), added ProviderModelByMiniMax struct
• pkg/model-provider/model_params_api.go - Registered MiniMax in ToModelParams(), added AppModelParamsMiniMax struct
• README.md / README_CN.md - Added MiniMax to supported providers list

Test plan

• 14 unit tests pass: Tags, TemperatureClamping, MaxTokensValidation, StreamOptions, ChatCompletionsUrl, JSONUnmarshal
• 3 integration tests pass (with MINIMAX_API_KEY): Unary, Stream, ToolCall
• go build ./... succeeds with no errors