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

{
#34301 is harder because it requires coordinating model-family registration, reasoning-mode detection, parameter validation, and regression coverage across multiple OpenAI variants; #26108 is comparatively localized to Gemini capability metadata and translation behavior.
}
https://github.com/berriai/litellm/issues/34301 2:1 https://github.com/berriai/litellm/issues/26108