Returns information about available AI models and their capabilities.

Use this endpoint to discover:

• Available model identifiers for use with /responses
• Model capabilities (vision, function calling, etc.)
• Token limits and other configuration

Optionally filter by model ID (UUID) using the litellm_model_id query parameter.

Errors: Unlike other AI Gateway routes, this operation does not return 401, 402, 403.
Documented error responses are 400 (invalid query input), 429, and 500.