BerriAI/litellm v1.71.2-nightly

on GitHub

What's Changed

• feat: Make gemini accept the openai parameter parallel_tool_calls by @aholmberg in #11125
• Fix #9295 docker-compose healthcheck test uses curl but curl is not in the image by @agajdosi in #9737
• [Feat] Add /image/edits support for Azure by @ishaan-jaff in #11160
• Fix deprecation_date value for llama groq models by @kiriloman in #11151
• [Fix] Rollback to httpx==0.27.0 by @ishaan-jaff in #11146
• Doc update for azure openai by @ketangangal in #11161
• Litellm fix GitHub action testing by @krrishdholakia in #11163
• [Feat - Contributor PR] Add Video support for Bedrock Converse by @ishaan-jaff in #11166
• [Fixes] Aiohttp transport fixes - add handling for aiohttp.ClientPayloadError and ssl_verification settings by @ishaan-jaff in #11162
• prevent leaking sensitive keys to langfuse + support forwarding /sso/key/generate to the server root path url by @krrishdholakia in #11165
• [Fix] - Duplicate maxTokens parameter being sent to Bedrock/Claude model with thinking by @ishaan-jaff in #11181
• Integration with Nebius AI Studio added by @Aktsvigun in #11143
• Codestral - return litellm latency overhead on /v1/completions + Add 'contains' support for ChatCompletionDeltaToolCall by @krrishdholakia in #10879
• Ollama Chat - parse tool calls on streaming by @krrishdholakia in #11171
• [Fix] Prometheus Metrics - Do not track end_user by default + expose flag to enable tracking end_user on prometheus by @ishaan-jaff in #11192
• [Fix]: Add cost tracking for image edits endpoint [OpenAI, Azure] by @ishaan-jaff in #11186
• VertexAI - codeExecution tool support + anyOf handling by @krrishdholakia in #11195
• Add Pangea provider to Guardrails hook by @ryanmeans in #10775
• Return anthropic thinking blocks on streaming + VertexAI Minor Fixes & Improvements (Thinking, Global regions, Parallel tool calling) by @krrishdholakia in #11194
• Azure OIDC provider improvements + OIDC audience bug fix by @nikoizs in #10054
• [Feat] Add well known MCP servers to LiteLLM by @ishaan-jaff in #11209
• Add missing request_kwargs to get_available_deployment call by @Nitro963 in #11202

New Contributors

• @agajdosi made their first contribution in #9737
• @ketangangal made their first contribution in #11161
• @Aktsvigun made their first contribution in #11143
• @ryanmeans made their first contribution in #10775
• @nikoizs made their first contribution in #10054
• @Nitro963 made their first contribution in #11202

Full Changelog: v1.71.1-nightly...v1.71.2-nightly

Docker Run LiteLLM Proxy

docker run \
-e STORE_MODEL_IN_DB=True \
-p 4000:4000 \
ghcr.io/berriai/litellm:main-v1.71.2-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