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

{
The right issue is harder because it requires careful low-level streaming/concurrency changes across synchronous provider integration, thread scheduling, buffering behavior, and regression testing without disrupting existing async and proxy paths. The left issue is comparatively contained UI work involving model-family detection, asset mapping, and table rendering.
}
https://github.com/berriai/litellm/issues/34502 5:1 https://github.com/berriai/litellm/issues/35095