What's Changed
- [Fix] UI - Revert to console outputs to reduce noise by @yuneng-jiang in #16981
- Permission Management - disable global guardrails by key/team by @krrishdholakia in #16983
- Model Armor - Logging guardrail response on llm responses by @krrishdholakia in #16977
- Add presidio pii masking tutorial with litellm by @krrishdholakia in #16969
- (fix) litellm_logging.py: fix mcp tool call response logging + (fix) responses_bridge: remove unmapped param error mid-stream - allows gpt-5 web search to work via responses api in
.completion()by @krrishdholakia in #16946 - [Docs] Model Compare UI by @yuneng-jiang in #16979
- fix(vertex_ai/image_generation_handler.py): fix default sample count … by @krrishdholakia in #16403
- [Fix] Revert UI - Organization Usage by @yuneng-jiang in #16980
- perf release notes for v1.80.5-stable by @AlexsanderHamir in #16978
- docs - ai gateway prompt management by @ishaan-jaff in #16990
- [Refactor] UI - Ensure Unique Keys in Navbar Menu Items by @yuneng-jiang in #16987
- feat: tool permission argument check by @uc4w6c in #16982
- [Refactor] UI - Minor Cosmetic Changes for Buttons, Add Notification for Delete Team by @yuneng-jiang in #16984
- fix: add nodejs and npm to docker image for prisma generate by @DmitriyAlergant in #16903
- chore: remove unused MCP_PROTOCOL_VERSION_HEADER_NAME constant by @uc4w6c in #17008
- feat: add backend support for OAuth2 auth_type registration via UI by @uc4w6c in #17006
- docs: cleanup README and improve agent guides by @Chesars in #17003
- feat: add UI support for registering MCP OAuth2 auth_type by @uc4w6c in #17007
- fix: Update broken documentation links in README by @Chesars in #17002
- [Fix]: (Vertex AI) handle global location in context caching by @prawaan in #16997
- fix: prevent duplicate spend logs in Responses API for non-OpenAI providers by @hula-la in #16992
- [Refactor] Change provider create fields to JSON by @yuneng-jiang in #16985
- [Fix] UI - Hide Default Team Settings From Proxy Admin Viewers by @yuneng-jiang in #16900
- OCI Provider: Fix pydantic validation errors during tool call with streaming. by @gotsysdba in #16899
- feat(proxy_server.py): preserve content field even if null by @krrishdholakia in #16988
- Update version and add preview tag by @krrishdholakia in #17032
- Document model pricing contribution process by @krrishdholakia in #17031
- Fix unspecified issue by @krrishdholakia in #17034
- Document event hook usage by @krrishdholakia in #17035
- Prompt security litellm by @lior-ps in #16365
- add claude opus 4.5 by @mubashir1osmani in #17043
- [Feat] New API - Claude Skills API (Anthropic) by @ishaan-jaff in #17042
- [Docs] Claude Skills API by @ishaan-jaff in #17052
- feat: add UI support for configuring tool permission guardrails by @uc4w6c in #17050
- add strands tutorial by @mubashir1osmani in #17039
- Add cost tracking for cohere embed passthrough endpoint by @Sameerlite in #17029
- fix: handle None or empty contents in Gemini token counter by @HaiyiMei in #17020
- Fix the azure auth format for videos by @Sameerlite in #17009
- Fix bedrock passthrough auth issue by @Sameerlite in #16879
- Add header forwarding in embeddings by @Sameerlite in #16869
- Integrate eleven labs text-to-speech by @Sameerlite in #16573
- Fix gpt-5.1 temperature support when reasoning_effort is "none" or not specified by @Sameerlite in #17011
- (fix) propagate x-litellm-model-id in responses by @raghav-stripe in #16986
- Fix: Distinguish permission errors from idempotent errors in Prisma migrations by @yuya2017 in #17064
- [Fix] Non Root Docker Build by @yuneng-jiang in #17060
- [Feature] Deleting a User From Team Deletes key User Ceated for Team by @yuneng-jiang in #17057
- Link to logging spec in callback docs by @krrishdholakia in #17049
- [Fix] Add aws_bedrock_runtime_endpoint into Credential Types by @yuneng-jiang in #17053
- Fix image edit endpoint by @emerzon in #17046
- Enhancement(helm): ServiceMonitor template rendering by @saar-win in #17038
- [Fix] UI - Add No Default Models for Team and User Settings by @yuneng-jiang in #17037
- fix(gemini): skip thinking config for image models by @Chesars in #17027
- Metadata: fix 401 when audio/transcriptions by @saar-win in #17023
- Make Bedrock image generation more consistent by @komarovd95 in #17021
- [Feature] Change Delete Modals to Common Component by @yuneng-jiang in #17068
- fix(vertex): fix CreateCachedContentRequest enum error by @sdip15fa in #16965
- merge main by @Sameerlite in #17082
- Add enforce user param functionality by @krrishdholakia in #17088
- Fix
reasoning_effort="none"not working on Azure for GPT-5.1 by @Sameerlite in #17071 - Add search API logging and cost tracking in LiteLLM Proxy by @Sameerlite in #17078
- Added support for azure anthopic models via chat completion by @Sameerlite in #16886
- Add vertex ai image gen support for both gemini and imagen models by @Sameerlite in #17070
New Contributors
- @prawaan made their first contribution in #16997
- @lior-ps made their first contribution in #16365
- @HaiyiMei made their first contribution in #17020
- @yuya2017 made their first contribution in #17064
- @saar-win made their first contribution in #17038
- @sdip15fa made their first contribution in #16965
Full Changelog: v1.80.5-nightly...v1.80.5.dev1