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

{
The left issue is harder because it lacks a reliable reproduction and may require tracing cross-provider request translation, tool-call normalization, and model-specific validation. The right issue has a deterministic response-conversion path and a narrower fix with straightforward regression coverage.
}
https://github.com/berriai/litellm/issues/30515 3:1 https://github.com/berriai/litellm/issues/27492