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

{
The right issue spans authorization-sensitive routing, multiple synchronous code paths, and cache behavior, requiring careful regression testing across router and cache layers. The left issue is more localized to token accounting and request parsing, with a narrower validation surface.
}
https://github.com/berriai/litellm/issues/31260 5:3 https://github.com/berriai/litellm/issues/32865