- Introduced new constants for Gemini models endpoint, cache TTL, and cache structure.
- Implemented `get_gemini_models` function to fetch models from the Gemini API, with caching and error handling.
- Added `get_gemini_models_error` function to retrieve error messages related to Gemini model fetching.
- Updated `get_models_by_provider` to include Gemini models retrieval.
- Enhanced model options retrieval to include Gemini models.
- Added `to_gemini_message` method in `LLMMessage` class for formatting messages compatible with Gemini API.
- Created `GeminiApi` class for managing Gemini API interactions.
Updated the README to include OpenRouter support alongside GPT and Claude. Added functionality in `llm_aio.py` to fetch and cache models from OpenRouter and OpenAI, improving model retrieval efficiency. This change enhances user experience by providing live model updates and reducing API calls.
oshtz