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

{
The left requires coordinated dashboard data interpretation, conditional presentation, and regression coverage across token/cost paths; the right is a localized backend classification fix with focused tests.
}
https://github.com/berriai/litellm/issues/32045 3:1 https://github.com/berriai/litellm/issues/28163