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

{
The right issue requires modifying a hot-path caching condition, handling call-type configuration safely, and adding regression coverage; the left is a narrowly scoped model metadata update.
}
https://github.com/berriai/litellm/issues/31862 8:1 https://github.com/berriai/litellm/issues/28307