feat: add MiniMax M3 prompt driver

[octo-patch]

Summary

Add MiniMax as a new LLM provider, compatible with the existing OpenAI-compatible provider architecture.

Changes

• Add MinimaxPromptDriver (subclasses OpenAiChatPromptDriver, points base_url at MiniMax's OpenAI-compatible endpoint), mirroring GrokPromptDriver
• Add MinimaxTokenizer with MiniMax's 1M-token context window
• Register both in griptape/drivers/__init__.py and griptape/tokenizers/__init__.py
• Add unit tests for the driver (init, serialization round-trip) and tokenizer

Why

MiniMax's M-series models are served over an OpenAI-compatible API, so the driver reuses the existing OpenAiChatPromptDriver machinery just like GrokPromptDriver and PerplexityPromptDriver.

MinimaxTokenizer subclasses OpenAiTokenizer (tiktoken) rather than making a network call like GrokTokenizer, because MiniMax does not expose a dedicated public tokenization endpoint; token counting stays local and offline.

Testing

• pytest tests/unit/tokenizers/test_minimax_tokenizer.py tests/unit/drivers/prompt/test_minimax_prompt_driver.py passing
• ruff format --check and ruff check clean

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!