BerriAI/litellm v1.77.5.rc.1

on GitHub

What's Changed

• Add supported text field to anthropic citation response by @TomeHirata in #14164
• fix: SSO "Clear" button writes empty values instead of removing SSO config by @Jetemple in #14826
• Fix: make fastuuid an optional dependency for proxy, graceful fallback to stdlib uuid by @akshoop in #14818
• Fix: support claude code auth via subscription (anthropic) by @hazyone in #14821
• Fix concurrency/scaling when many Python threads do streaming using sync completions by @leventov in #14816
• Fix load credentials in token counter proxy by @eycjur in #14808
• Adding langfuse usage details for cached tokens by @fabriciojoc in #10955
• doc: make the README document clearer by @mrFranklin in #14860
• feat: New model - Add support for Qwen models family & Deepseek 3.1 to Amazon Bedrock by @onlylonly in #14845
• Added vertex_ai/qwen models and azure/gpt-5-codex by @mubashir1osmani in #14844
• Add sambanova deepseek v3.1 and gpt-oss-120b by @luisfucros in #14866
• [Fix] LakeraAI v2 Guardrail - Ensure exception is raised correctly by @ishaan-jaff in #14867
• Add /user/list to management routes by @ishaan-jaff in #14868
• [Memory Leak Fix] Fix InMemoryCache unbounded growth when TTLs are set by @Copilot in #14869
• Bump tar-fs from 3.0.10 to 3.1.1 in /docs/my-website by @dependabot[bot] in #14872
• Fix: make pondpond as optional dependency for proxy extras, disable object pooling gracefully by @akshoop in #14863
• Revert "Fix: make pondpond as optional dependency for proxy extras, disable object pooling gracefully" by @ishaan-jaff in #14880
• [Feat] Initial support for scheduled key rotations by @ishaan-jaff in #14877
• Fix: Support custom entity types in Presidio guardrail with Union[PiiEntityType, str] by @arsh72 in #14899
• Remove useful links from admin settings by @ishaan-jaff in #14918
• Update docs for session management availability by @berri-teddy in #14914
• [Feat] Logging - datadog callback Log message content w/o sending to datadog by @ishaan-jaff in #14909
• fix: use fastuuid helper by @AlexsanderHamir in #14903
• [Feat] - Cost Tracking - show input, output, tool call cost breakdown in StandardLoggingPayload by @ishaan-jaff in #14921
• Corrected docs updates sept 2025 by @berri-teddy in #14916
• (Feat) Add BitBucket Integration for Prompt Management by @Sameerlite in #14882
• Remove aiohttp_ prefix from config by @berri-teddy in #14920
• add noma guardrail provider to ui by @vpbill in #14415
• fix: prisma client state retries by @mubashir1osmani in #14925
• 🐛 Fix a bug where openai image edit siltently ignores multiple images by @kgritesh in #14893
• feat: improve opik integration code by @mrFranklin in #14888
• Add gpt-5 and gpt-5-codex to OpenRouter cost map by @huangyafei in #14879
• GPT-3.5-Turbo price updated. by @oytunkutrup1 in #14858
• Fix: revert fastuuid optional dependency, always use fastuuid in .__uid helper by @akshoop in #14941
• [Feat] Add support for Gemini 2.5 Flash and Flash-lite preview models (09-2025 release) by @ishaan-jaff in #14948
• docs: add documentation for additional cost-related keys in custom pricing by @thiswillbeyourgithub in #14949
• [Feat] Add new anthropic web fetch tool support by @ishaan-jaff in #14951
• Revert incorrect changes to sonnet-4 max output tokens by @nherment in #14933
• fix: remove slow string operation by @AlexsanderHamir in #14955
• Azure Managed Batches - api key error by @mubashir1osmani in #14932
• fix mypy errors from bitbucket integration by @Sameerlite in #14959
• [Feat] UI - Allow scheduling key rotations when creating virtual keys by @ishaan-jaff in #14960
• fix: support documented params by @AlexsanderHamir in #14969
• [Fix] Parallel Request Limiter v3 - ensure Lua scripts can execute on redis cluster by @ishaan-jaff in #14968
• Error logging in SQS by @deepanshululla in #14974
• chore: setting devcontainer for develop by @22mSqRi in #14972
• Fix Anthropic streaming IDs (#14962) by @Maximgitman in #14965
• Fix inconsistent token configs for gpt-5 models by @danielmklein in #14942
• [Bug]: vLLM provider's rerank endpoint from /v1/rerank to /rerank by @subnet-dev in #14938
• Revert "Azure Managed Batches - api key error" by @krrishdholakia in #14978
• Add azure passthrough documentation by @berri-teddy in #14958
• stabilize main github tests by @krrishdholakia in #14979
• feat: initial commit for v2 oauth flow by @krrishdholakia in #14964
• test - fix mypy linting on ci/cd by @ishaan-jaff in #14981
• fix: reduce get_deployment cost to O(1) by @AlexsanderHamir in #14967
• fix: remove blocking create_task by @AlexsanderHamir in #14980
• Litellm ci cd linting fixes 09 29 2025 p2 by @krrishdholakia in #14982
• Revert "fix: remove blocking create_task" by @ishaan-jaff in #14984
• [Draft] v1.77.5 Release note by @ishaan-jaff in #14985

New Contributors

• @Jetemple made their first contribution in #14826
• @akshoop made their first contribution in #14818
• @hazyone made their first contribution in #14821
• @leventov made their first contribution in #14816
• @fabriciojoc made their first contribution in #10955
• @onlylonly made their first contribution in #14845
• @Copilot made their first contribution in #14869
• @arsh72 made their first contribution in #14899
• @berri-teddy made their first contribution in #14914
• @vpbill made their first contribution in #14415
• @kgritesh made their first contribution in #14893
• @oytunkutrup1 made their first contribution in #14858
• @nherment made their first contribution in #14933
• @deepanshululla made their first contribution in #14974

Full Changelog: v1.77.4-nightly...v1.77.5.rc.1

Docker Run LiteLLM Proxy

docker run \
-e STORE_MODEL_IN_DB=True \
-p 4000:4000 \
ghcr.io/berriai/litellm:main-v1.77.5.rc.1

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