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

{
#33955 is harder because it requires safely redesigning large-payload processing and concurrency behavior across batch workers, with substantial memory, performance, and regression risk. #25738 is narrower, primarily involving provider/model cost attribution for a streaming transport.
}
https://github.com/berriai/litellm/issues/33955 5:2 https://github.com/berriai/litellm/issues/25738