LiteLLM

This release focuses heavily on bug fixes across various providers (Anthropic, Gemini, Vertex AI) and internal proxy/UI stability, alongside introducing Docker image signing verification via cosign.

This release introduces Docker image signing verification instructions and includes numerous bug fixes across proxy, S3, logging, and guardrails. It also features a breaking change reducing the default Prometheus latency histogram buckets.

This release introduces Docker image signature verification using cosign, adds new API endpoints and features like Ramp callbacks, and includes numerous bug fixes across routing, cost calculation, and Docker builds.

This release introduces cryptographic signing for all LiteLLM Docker images using cosign to improve supply chain security. Users are encouraged to verify image authenticity using the provided signing instructions.

This release introduces cryptographic signing for all LiteLLM Docker images using cosign to improve supply chain security. Instructions are provided for verifying image signatures.

This release introduces Docker image signature verification via cosign and adds significant features for team management, including project-level guardrails and enhanced UI controls for keys and rate limits. Numerous bug fixes address issues across the UI, proxy, and agent integrations.

This patch release introduces new UI features for rate limiting and credential exposure, adds embedding usage tracking fallback, and includes several UI and Docker build bug fixes. It also details how to verify the integrity of LiteLLM Docker images using cosign.

This release introduces Docker image signing verification using cosign and adds significant new features to the router, including order-based fallback and health-check-driven routing. Several routing and Docker build fixes were also implemented.

This release introduces Docker image signing using cosign for improved security verification. It also includes minor version bumps for internal components.

This release introduces Docker image signing verification via cosign and adds significant features to the UI and Proxy, including project-level guardrails and Azure Entra ID credential support. Several bug fixes address issues related to UI filtering, team management, and Docker image loading.

This release introduces Docker image signing using cosign to ensure image integrity. Users can now verify the authenticity of LiteLLM Docker images before deployment.

This release introduces Docker image signing using cosign to ensure image integrity. Users can now verify the authenticity of LiteLLM Docker images before deployment.

This was a test release from the LiteLLM team to test a new signing process. Co-sign verification is temporarily non-functional.

This release focuses heavily on infrastructure hardening, security improvements (like OpenSSF Scorecard adoption and dependency pinning), and numerous bug fixes across providers like Gemini, Anthropic, and Azure OpenAI. New features include Prometheus metrics for batch operations and budget enforcement across multi-pod deployments.

This release focuses heavily on UI enhancements, including new organizational and team management features, alongside critical bug fixes related to client handling, logging redaction, and database migration stability. A new provider for Amazon Nova models on SageMaker was also introduced.

This release introduces enhancements to Prometheus metrics and proxy callback tracking, alongside a fix for missing post-call guardrail logs in the proxy.

This release appears to be a release candidate (v1.82.6.rc.2) with no detailed changes provided in the snippet, linking only to the comparison view.

This release introduces support for the sagemaker_nova provider and includes numerous UI enhancements related to key management and team settings. Several critical bug fixes address issues with client eviction, logging redaction, and database migration stability.

This release contains internal development changes between versions v1.82.5.dev.1 and v1.82.3.dev.5.

This release focuses heavily on bug fixes and stability improvements across various providers like Anthropic, Gemini, and Vertex AI, alongside introducing new features such as Akto Guardrails integration and prompt management support for the Responses API.

This release appears to be a development release (v1.82.6.dev1) following a release candidate (v1.82.6.rc.1), with the full changelog available via the provided comparison link.

This release introduces new features like the sagemaker_nova provider and various UI enhancements, alongside numerous bug fixes across providers (Anthropic, Gemini, Vertex AI) and infrastructure improvements, including security updates and better logging redaction.

This release focuses heavily on UI modernization, especially around Teams and MCP management, alongside numerous bug fixes for proxy guardrails, logging, and key management endpoints. It also introduces infrastructure improvements and new environment variable support for Anthropic.

This release focuses heavily on stability and feature parity across various providers, including significant fixes for Anthropic reasoning summaries, Gemini image handling, and Vertex AI pricing/streaming. It also introduces new features like Akto Guardrails integration and control plane management.

This development release primarily focuses on fixing bugs related to parameter passing for Azure OpenAI and improving streaming response handling across various providers.

This release focuses heavily on UI modernization, infrastructure improvements including a control plane for worker management, and numerous fixes across various providers like Bedrock, Gemini, and Moonshot. Several documentation updates regarding pricing and caching were also included.

This release focuses heavily on UI modernization, especially around Teams features, and includes numerous fixes across various providers like Bedrock, Gemini, and Anthropic. A new control plane for multi-proxy worker management was also introduced.

This release focuses heavily on UI enhancements, including new organizational and team management features, alongside critical bug fixes related to logging redaction, database migrations, and dependency security updates.

This release focuses heavily on bug fixes across the proxy, UI, and logging systems, alongside infrastructure improvements like black formatting and dependency updates. Key features include enhanced Anthropic integration and new control plane capabilities for multi-proxy management.

This release focuses heavily on UI enhancements, including new organizational and team management features, alongside numerous bug fixes addressing security, logging redaction, and database migration stability. A new provider, sagemaker_nova, was also introduced.

This release focuses heavily on bug fixes across the proxy, UI, and logging systems, alongside infrastructure updates like code formatting and dependency regeneration. New features include enhanced team endpoint access control and improved Anthropic environment variable support.

This release focuses heavily on bug fixes across proxy functionality, UI components, and provider integrations, alongside infrastructure updates like dependency regeneration and code formatting. New features include enhanced team listing access control and proxy-wide rate limits.

This development release primarily focuses on fixing parameter passing issues for the Azure OpenAI provider, specifically for max_tokens and temperature.

This release focuses heavily on UI enhancements, including new organizational controls and modernized settings pages. It also includes numerous bug fixes addressing security issues, database migration stability, and provider integration reliability.

This release is a stable patch for Litellm version 1820, implemented by @shivamrawat1.

This release focuses heavily on UI enhancements, including organization and team management features, alongside critical bug fixes related to client handling, logging redaction, and database migration stability. A new provider for Amazon Nova models on SageMaker was also introduced.

This release focuses heavily on UI enhancements, including organization and team management features, alongside critical bug fixes related to logging redaction, database migrations, and dependency security updates.

This patch release focuses heavily on bug fixes across various integrations (Anthropic, Azure, HuggingFace) and infrastructure stability, alongside introducing a new User Info V2 Endpoint and allowing organization_id setting on key updates.

This release focuses heavily on bug fixes, infrastructure improvements, and significant updates to the UI, including new endpoints and administrative features. Key fixes address security vulnerabilities, proxy stability, and various provider integrations.

This release introduces several new UI features, including user info endpoints and team management tools, alongside numerous bug fixes addressing issues in proxy stability, cost tracking, and API integrations. Dependency updates for security were also performed.

This release primarily consists of internal development updates, as indicated by the nightly/dev tag, with a link provided to the detailed comparison.

This release focuses heavily on enhancing Multi-Cloud Proxy (MCP) functionality, including new authentication methods, BYOM support, and UI improvements. It also brings numerous bug fixes across providers like Gemini, Anthropic, and OpenAI, alongside new feature support for transcription and image editing.

This release focuses heavily on expanding support for various models (Gemini, Mistral Voxtral, Qwen3.5) and enhancing the Multi-Cloud Proxy (MCP) features, including new authentication and UI workflows. Numerous bug fixes address streaming issues, provider specific mappings, and security concerns.

This release appears to be a development build (v1.82.0.dev6) referencing the comparison against the previous stable release (v1.82.0-stable), but the specific changes are not detailed in the provided notes.

This release introduces significant feature enhancements, particularly around MCP server support, new model integrations (Voxtral, Qwen3.5), and expanded guardrail capabilities. Numerous bug fixes address issues across streaming, Anthropic integration, and UI stability.

This release primarily points to the full changelog for version v1.82.0.dev4 compared to v1.82.0-stable.

This release appears to be a development build (v1.82.0.dev3) following v1.82.0-stable, with the full changelog available via the provided comparison link.

This release focuses primarily on bug fixes across various components, including routing, caching, provider integrations (Fireworks, Sagemaker, Vertex AI, Bedrock), and the Responses API streaming bridge. Several provider-specific URL and request body issues were resolved.

This release corresponds to a nightly update identified by the comparison link between v1.82.1-nightly.1 and v1.82.1-nightly.1-re.

This release focuses heavily on bug fixes across numerous providers (Gemini, Anthropic, OpenAI, Azure, etc.), improves streaming performance, and introduces new features like UI Projects and enhanced model cost tracking. A key structural change involves updating response types to use ModelResponseStream instead of StreamingChoices.

This stable release primarily addresses a bug where HTTP/SDK clients were incorrectly closed during LLMClientCache eviction.

This release primarily consists of internal updates and bug fixes, referenced by the full changelog link.

This release focuses heavily on bug fixes across numerous providers (Gemini, Anthropic, OpenAI, Azure, etc.), improves streaming performance, and introduces new features like UI Projects and enhanced model cost tracking. It also deprecates older xAI models.

This release primarily consists of internal updates and links to the detailed comparison between nightly builds.

This release focuses heavily on bug fixes across various integrations, including completion bridges, routing, proxy endpoints, caching, and specific provider integrations like Vertex AI, Fireworks, Sagemaker, and Bedrock. New features include silent metric recording and helmchart deployment strategy support.

This patch release primarily focuses on bug fixes, specifically addressing incorrect parameter passing for Azure OpenAI and improving streaming response handling across providers.

This release focuses heavily on bug fixes across various components including routing, caching, proxy endpoints, and specific provider integrations like Fireworks, Sagemaker, and Vertex AI. Key improvements involve error handling during retries and data preservation in streaming responses.

This development release focuses on fixing bugs related to parameter passing for Azure OpenAI and improving streaming response handling across various providers.

This release focuses heavily on bug fixes across numerous providers (Gemini, Anthropic, OpenAI, Azure, etc.), improves streaming performance, and introduces new features like UI Projects and enhanced model cost tracking. A key change involves updating response types to use ModelResponseStream instead of StreamingChoices.

This release focuses heavily on bug fixes across various providers (Gemini, Anthropic, OpenAI, Azure AI, etc.), enhances model cost mapping, and introduces new features like UI Projects and improved streaming latency. It also deprecates older xAI models.

This release focuses on fixing critical parameter passing issues for the Azure OpenAI provider, specifically for max_tokens and timeout settings.

This release contains a patch update for the v1.81.14-stable.gpt-5.4 branch, with details available in the full changelog link.

This release corresponds to release candidate 2 (v1.82.rc.2) following the v1.82.0-nightly build.

This patch release primarily focuses on fixing an issue related to Azure OpenAI API key passing and resolving a streaming max_tokens bug.

This patch release focuses on fixing critical bugs related to tool call handling for the Gemini 1.5 Flash model on the completion endpoint.

This release focuses heavily on bug fixes across numerous providers (Gemini, Anthropic, OpenAI, Azure, etc.), enhances model tracking and cost mapping, and introduces performance improvements for streaming latency. It also adds new UI features for project management and access control.

This release focuses heavily on bug fixes across various providers (Gemini, Anthropic, OpenAI, Azure, OpenRouter) and internal systems like MCP and caching. It also introduces new model support, performance optimizations for spendlogs and streaming latency, and new UI features for project management.

This release corresponds to a patch update in the v1.81.14-stable.gpt-5.4 series, detailed in the linked comparison.

This patch release primarily addresses a bug related to the incorrect passing of the `max_tokens` parameter to the OpenAI API for specific models.

This release focuses heavily on bug fixes across numerous providers (Gemini, Anthropic, OpenAI, Azure, etc.), improves streaming performance, and introduces new features like UI Projects and enhanced model cost tracking.

This patch release primarily addresses bugs related to the Azure OpenAI provider, ensuring correct handling of completion and streaming responses.

This patch release primarily focuses on fixing bugs related to parameter passing for Azure OpenAI and improving streaming response handling.

This patch release addresses critical bugs related to Azure OpenAI API key passing and streaming response handling across various providers.

This release corresponds to version v1.81.14.gemini-batches, with detailed changes available via the comparison link.

This release contains a full changelog available via the provided comparison link between v1.81.14-stable and v1.81.14-redirect-alpha.

This release introduces significant UI features related to Projects and Access Groups, adds several new models and cost map entries, and includes numerous performance optimizations and bug fixes across proxy, caching, and model handling.

This release includes a full changelog available via the provided comparison link, marking the stable release v1.81.14.

This release focuses heavily on UI enhancements, particularly around Projects management and access control (RBAC), alongside numerous performance optimizations and bug fixes across model integrations and proxy functionality. Key additions include new model support and improved tracing capabilities.

This release contains minor updates and bug fixes, detailed in the linked comparison between v1.81.6.rc.6 and v1.81.6.rc.9.

This patch updates the Anthropic beta headers configuration to enable the 1M context window feature for Anthropic and Vertex AI, and adds extensive beta header mappings for the Databricks provider.

This release introduces significant feature enhancements, particularly around UI projects, access control, and model support, alongside numerous bug fixes and performance optimizations across various integrations and internal systems.

This release primarily focuses on bug fixes, specifically addressing issues with passing `max_tokens` to Azure OpenAI and improving streaming response handling across providers.

This release includes minor updates and fixes between release candidates v1.81.6.rc.4 and v1.81.6.rc.6.

This release contains a series of minor updates and fixes between release candidates v1.81.6.rc.5 and v1.81.14.rc.2.

This release focuses on stability and UI improvements, introducing pagination for key aliases, fixing OOM issues, and enhancing real-time WebSocket error handling. A previously added /public/supported_endpoints feature was reverted.

This release focuses heavily on security fixes, performance optimizations across various components, and significant feature additions to the proxy and UI, particularly around tool policies, agent key assignment, and real-time guardrails. Critical security vulnerabilities were also addressed.

This release focuses heavily on security fixes, performance optimizations across various components, and significant feature additions to the proxy, guardrails, and UI, including support for new models and enhanced logging/tracing.

This release focuses on bug fixes, specifically removing a duplicate schema definition and introducing spend logging within the Release Candidate build.

LiteLLM v1.81.6.rc.4 focuses on stability by fixing several bugs related to parameter passing (like max_tokens for Azure) and response streaming across various providers.

This patch release backports a stateless fix for StreamableHTTPSessionManager to the v1.81.12-stable branch.

This release appears to be a maintenance or internal version bump, indicated by the comparison range v1.81.12-stable to v1.81.12-stable_gpt-5.3, with no detailed changes provided in the snippet.

This release provides an update to the nightly build, linking to the full comparison between the previous development version and the new nightly version.

This release focuses primarily on bug fixes across various integrations, including recovery from zombie processes, improved pricing application, and fixes related to rate limiting and authentication flows. Several recent fixes were also reverted.

This release addresses a critical CVE by fixing issues related to the identification and handling of the 'presidio' model integration.

This release primarily contains internal updates and links to the full comparison view for detailed changes.

This release introduces new features like custom favicon support, configurable semantic cache dimensions, and guardrail policy versioning. Numerous bug fixes address issues across Bedrock, budget configuration, testing stability, and model cost mapping.

This release introduces new model pricing, enhanced content filtering with topic blockers, and a new guardrail hook for voice transcription, alongside a fix for Replicate server communication.

This release introduces several new features, including custom favicon support, configurable semantic cache dimensions, and new guardrail policies. Numerous bug fixes address issues across Bedrock integration, budget calculation, testing stability, and data normalization.

This release introduces new features like custom proxy favicons, configurable semantic cache dimensions, and guardrail policy versioning. Numerous bug fixes address issues across Bedrock integration, budget tracking, testing stability, and data normalization.

This release introduces new features like custom favicon support for the proxy, configurable semantic cache dimensions, and guardrail monitoring. It also includes numerous bug fixes across providers like Bedrock and Anthropic, and utility improvements like parameter key normalization.