v1.89.2 - Cost Tracking & Model-List Fixes

Update: no performance regression found

An earlier version of this note flagged a potential throughput regression. We investigated and could not confirm or reproduce any regression in the released version. The one report we received came from a deployment running custom code on top of what we shipped, and our testing points to those changes, not LiteLLM, as the likely cause.

Correctness and error rates were never affected. If you're on this version, there's nothing you need to do.

We're still monitoring incoming reports and will update this note if anything changes.

Deploy this version

Docker

docker run \
-e STORE_MODEL_IN_DB=True \
-p 4000:4000 \
docker.litellm.ai/berriai/litellm:1.89.2

Pip

pip install litellm==1.89.2

v1.89.2 is a patch release on top of v1.89.1. It hardens cost tracking around service_tier, corrects /v1/models listing for team and BYOK setups, and tightens vector-store access and OTEL error reporting.

What's Changed

• fix(cost): stop a non-string service_tier from silently dropping cost tracking - PR #30690
• fix(anthropic): price and surface the response service_tier in cost tracking - PR #30558
• fix(proxy): list the public team model name in /v1/models - PR #30588
• feat(proxy): add an opt-in healthy_only filter to GET /v1/models - PR #30130
• fix(proxy): resolve list-files credentials from team BYOK deployments - PR #30495
• fix(proxy): allow internal roles to access vector store CRUD routes - PR #30503
• fix(otel): record the full error message on the standard exception event in OTEL v2 - PR #30380

Full Changelog

https://github.com/BerriAI/litellm/compare/v1.89.1...v1.89.2