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

{
The left task is harder because it requires diagnosing an elusive latency regression across provider translation, streaming behavior, authentication, retries, and deployment boundaries. The right task is a broader feature integration, but its interfaces and validation flow can be designed and tested more directly.
}
https://github.com/berriai/litellm/issues/26699 3:2 https://github.com/berriai/litellm/issues/26514