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

{
The right issue is harder because it spans pricing-model fallback construction, cache-aware billing behavior, tiered field propagation, and regression coverage across router and cost-calculation paths. The left issue is comparatively localized to normalizing a reasoning value consistently between adjacent request transformations.
}
https://github.com/berriai/litellm/issues/26806 3:1 https://github.com/berriai/litellm/issues/28196