v1.84.5 - Azure AD, Batch Auth & Passthrough Backports

Deploy this version

Docker

docker run \
-e STORE_MODEL_IN_DB=True \
-p 4000:4000 \
docker.litellm.ai/berriai/litellm:1.84.5

Pip

pip install litellm==1.84.5

v1.84.5 is a patch release on top of v1.84.4. It backports six staged fixes covering Azure AD token refresh, batch and video model routing, org-scoped team key creation, Vertex Claude effort handling, and duplicate passthrough cost callbacks.

What's Changed

• fix(azure): preserve AD token refresh in the v1 OpenAI client path - PR #28627
• fix(proxy): map a stripped batch body.model back to the proxy alias so key access checks pass - PR #29264
• fix(proxy): resolve managed video model ids through the router before auth, budget, and key checks - PR #29545
• fix(key_generate): let team members create keys on org-scoped teams (regression since v1.84.0-rc.1) - PR #29310
• fix(vertex): strip output_config.effort for Vertex Claude models that reject it, such as Haiku 4.5 - PR #29585
• fix(passthrough): stop duplicate cost callbacks for Anthropic streaming pass-through - PR #29598

Full Changelog

https://github.com/BerriAI/litellm/compare/v1.84.4...v1.84.5