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

{
The left issue is harder because it spans shared Pydantic model composition, provider-specific configuration semantics, initialization flow, and compatibility testing. The right issue is comparatively localized to a transformation branch with focused regression coverage, though it still carries some API-compatibility risk.
}
https://github.com/berriai/litellm/issues/32061 3:1 https://github.com/berriai/litellm/issues/34978