BerriAI/litellm v1.72.7-nightly

on GitHub

What's Changed

• feat(azure): Make Azure AD scope configurable by @kjoth in #11621
• Litellm stable docs 06 14 2025 p2 by @krrishdholakia in #11738
• Release note updates + Responses API Bridge improvements by @krrishdholakia in #11740
• VertexAI Anthropic - streaming passthrough cost tracking by @krrishdholakia in #11734
• Fix PrometheusLogger label_filters initialization for non-premium users by @colesmcintosh in #11764
• Add Vertex Imagen-4 models by @emerzon in #11767
• Users page buttons repositioned by @NANDINI-star in #11771
• #11748: Added Mistral Small to BEDROCK_CONVERSE_MODELS for Converse A… by @shagunb-acn in #11760
• [Security] Fixes for docs by @ishaan-jaff in #11776
• [Security] - Add Trivy Security Scan for UI + Docs folder - remove all vulnerabilities by @ishaan-jaff in #11778
• [Fix] Pass through - Langfuse don't log request to Langfuse passthrough on Langfuse by @ishaan-jaff in #11768
• [Deps] Fix aiohttp version requirement by @ishaan-jaff in #11777
• AWS credentials no longer mandatory by @MadsRC in #11765
• build(deps): bump next from 14.2.26 to 14.2.30 in /ui/litellm-dashboard by @dependabot in #11720
• feat: update the feature of ollama_embeddings to work on a sync api by @Abiji-2020 in #11746
• [Feat] Day-0 Support for OpenAI Re-usable prompts Responses API by @ishaan-jaff in #11782
• SSO - Allow passing additional headers + Spend Tags - automatically track spend by user agent (allows cost tracking for claude code) by @krrishdholakia in #11781
• JWT Auth - correctly return user email + UI Model Update - Allow editing model access group for existing model by @krrishdholakia in #11783
• Allow /models to return correct models for custom wildcard prefixes by @krrishdholakia in #11784
• Fix JSX syntax error in documentation causing Vercel deployment failure by @colesmcintosh in #11818
• [Fix] Bug Fix for using prom metrics config by @ishaan-jaff in #11779
• [Bug Fixes] MCP - using MCPs defined on config.yaml + fix for MCP error Team doesn't exist in cache by @ishaan-jaff in #11824
• new gemini model pricing + a few openrouter models model_prices_and_context_window.json by @salzubi401 in #11803
• Update bedrock guardrail docs by @orolega in #11826
• [Feat] v2 Pass through endpoints - Add support for subroutes for pass through endpoints + Cleaned up UI by @ishaan-jaff in #11827
• Fix vertex ai claude thinking params by @X4tar in #11796
• Implement health check backend API and storage functionality - fix ci/cd by @colesmcintosh in #11852
• [Fix] v1/messages endpoint always uses us-central1 with vertex_ai-anthropic models by @ishaan-jaff in #11831
• Fix #11856: Update billing.md docs to call the new GPT-4o model by @karen-veigas in #11858
• Add LiteLLM_HealthCheckTable to database schema by @colesmcintosh in #11677
• [SCIM] Add Error handling for existing user on SCIM by @ishaan-jaff in #11862
• feat(speech/): working gemini tts support via openai's /v1/speech endpoint by @krrishdholakia in #11832
• Completion-To-Responses Bridge: Support passing image url's by @krrishdholakia in #11833
• Implement health check frontend UI components and dashboard integration by @colesmcintosh in #11679
• Remove retired version of gpt-3.5 from prometheus.md by @Shankyg in #11859
• Minor Fixes by @krrishdholakia in #11868
• Fix boto3 tracer wrapping for observability by @colesmcintosh in #11869
• [Feat] Passthrough - Add support for setting custom cost per pass through request by @ishaan-jaff in #11870
• [Fix] SCIM - Add SCIM PATCH and PUT Ops for Users by @ishaan-jaff in #11863
• [UI] - Move passthrough endpoints under Models + Endpoints by @ishaan-jaff in #11871
• Fix gemini 2.5 flash config by @lowjiansheng in #11830
• Fix: #11853 Updated model version in alerting.md for latest model called when adding metadata to proxy calls. by @karen-veigas in #11855
• [Bug Fix] - Ensure "Request" is tracked for pass through requests on LiteLLM Proxy by @ishaan-jaff in #11873
• Add user agent tags in spend logs payload + Fix Azure ai content type + Fix passing dynamic credentials on retrieve batch by @krrishdholakia in #11872
• UI - allow setting default team for new users by @krrishdholakia in #11874
• Revert "UI - allow setting default team for new users" by @krrishdholakia in #11876
• Revert "Revert "UI - allow setting default team for new users"" by @krrishdholakia in #11877
• Fix default team settings by @NANDINI-star in #11887
• [Feat] UI - Add Allowed MCPs to Creating/Editing Organizations by @ishaan-jaff in #11893
• [Feat] Enable Tool Calling for meta_llama by @ishaan-jaff in #11895
• fix(vertex_ai): Handle missing tokenCount in promptTokensDetails (#11… by @ishaan-jaff in #11896
• [Bug Fix]: Fix gemini - web search error with responses API by @ishaan-jaff in #11894
• Revert "Users page buttons repositioned" by @krrishdholakia in #11904
• [Feat] V2 Add Pass through endpoints on UI by @ishaan-jaff in #11905
• Fix clickable model ID in health check table by @colesmcintosh in #11898
• Fix health check UI table design by @colesmcintosh in #11897
• [Bug Fix] add missing flash-2.5-flash-lite for gemini provider, fix gemini-2.5-flash pricing by @fcakyon in #11901
• feat: add workload identity federation between GCP and AWS by @pascallim in #10210

New Contributors

• @kjoth made their first contribution in #11621
• @shagunb-acn made their first contribution in #11760
• @MadsRC made their first contribution in #11765
• @Abiji-2020 made their first contribution in #11746
• @salzubi401 made their first contribution in #11803
• @orolega made their first contribution in #11826
• @X4tar made their first contribution in #11796
• @karen-veigas made their first contribution in #11858
• @Shankyg made their first contribution in #11859
• @pascallim made their first contribution in #10210

Full Changelog: v1.72.6.dev1...v1.72.7-nightly

Docker Run LiteLLM Proxy

docker run \
-e STORE_MODEL_IN_DB=True \
-p 4000:4000 \
ghcr.io/berriai/litellm:main-v1.72.7-nightly

Don't want to maintain your internal proxy? get in touch 🎉

Hosted Proxy Alpha: https://calendly.com/d/4mp-gd3-k5k/litellm-1-1-onboarding-chat

Load Test LiteLLM Proxy Results