Sorting LiteLLM GitHub issues by difficulty to implement. Higher = harder.

{
The left issue is harder because it likely requires tracing authorization, model-list retrieval, and proxy UI behavior across multiple layers, with regression and role-specific testing. The right issue is comparatively localized to Gemini request construction and authentication handling, with a narrower provider-specific test surface.
}
https://github.com/berriai/litellm/issues/17475 3:1 https://github.com/berriai/litellm/issues/34872