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

{
The left requires deeper provider-specific translation/debugging across multimodal and tool-message paths, with greater compatibility and regression risk. The right is a more localized exception-handling and fallback adjustment with a narrower test surface.
}
https://github.com/berriai/litellm/issues/35236 3:1 https://github.com/berriai/litellm/issues/34243