litellm Releases

A major release focused on security, stability, and UI improvements. Key updates include Docker image signing, dark mode UI support, and robust proxy/router stability fixes. It also introduces support for new providers like Gemini 3.5 transcribe and Vertex AI Interactions API.

Sep 6, 2026

This is a Docker-only release without a PyPI package, featuring an OpenTelemetry v2 fix that adds cache token counts to spans to correct spend under-counting. The rolling `latest` and `main-stable` image tags have been updated to point to this version.

Sep 2, 2026

A maintenance patch for the 1.97 line that refreshes base-image and dependency pins to address known CVEs without altering LiteLLM behavior. The rolling `latest` and `main-stable` tags are intentionally not moved by this release.

Sep 2, 2026

A major release featuring Docker image signature verification, a comprehensive UI refactor from Ant Design to shadcn/ui and React 19, and the addition of new providers like Amazon Comprehend Medical and Valkey. It includes numerous bug fixes for Bedrock, Azure, Batches, and Guardrails, as well as new features like async Rust OCR and complexity router enhancements.

Sep 1, 2026

This release introduces comprehensive Provisioned Throughput (PTU) configuration capabilities, adds vector store index listing, and enhances rate limiting with configurable estimated output tokens.

Aug 23, 2026

A feature-rich release adding Cursor thinking support, guardrails integration, and complexity router session affinity, alongside improvements to the playground, spend dashboard, and OpenTelemetry metrics.

Aug 16, 2026

A maintenance release that backports critical proxy request-handling improvements and refreshes runtime dependencies to address a previous PyPI publishing failure.

Aug 11, 2026

A maintenance release focused on backporting proxy request-handling improvements and refreshing runtime dependencies to ensure stability.

Aug 11, 2026

A maintenance release that backports specific proxy fixes and dependency updates to the 1.94.x series.

Aug 11, 2026

A maintenance release for the 1.93.x stable line focusing on proxy request-handling fixes and runtime dependency refresh.

Aug 11, 2026

A maintenance release for the 1.92.x stable line involving proxy request-handling maintenance and runtime dependency refresh.

Aug 11, 2026

A maintenance release backporting proxy request-handling improvements to the stable branch.

Aug 11, 2026

A maintenance release for the 1.90.x stable line featuring backports of proxy request-handling fixes and runtime dependency refresh.

Aug 11, 2026

A maintenance release for the 1.89.x stable line that backports multiple proxy request-handling and maintenance fixes.

Aug 11, 2026

A maintenance release for the 1.89.x stable branch. This version includes a backport of a release chore and a revert of a GitHub Actions sync issue.

Aug 11, 2026

A maintenance release for the 1.88.x stable line focusing on proxy request-handling fixes and runtime dependency refresh.

Aug 11, 2026

This release introduces Docker image signing for security verification, enhances Prometheus metrics with service tier labels, and includes numerous fixes for proxy routing, authentication, and UI improvements.

Aug 10, 2026

A maintenance release for the 1.94.x stable line that backports proxy request-handling fixes and refreshes runtime dependencies.

Aug 8, 2026

A beta release for Unsloth focusing on new model support (DeepSeek V4 Flash and Kimi K3), improved download stability, and various bug fixes across the Studio and Desktop applications.

Aug 4, 2026

A major feature release introducing Rust backend enhancements (Anthropic /messages routing, OpenAI Responses API), a new Cost Optimization UI page, and new guardrails (Compresr, Straiker). Adds Claude Opus 5 and Gemini 3.6/3.5 pricing.

Aug 3, 2026