LiteLLM has announced the release of v1.65.0-stable, bringing significant advancements to its platform. The highlight includes the addition of Model Context Protocol (MCP) support, allowing developers to centrally manage MCP servers integrated within LiteLLM. This enhancement provides an efficient way for developers to manage endpoints and utilize MCP tools, optimizing their workflow.
Another key update is the ability to view comprehensive usage analytics even after database logs exceed one million entries. This is made possible by a new scalable architecture that aggregates usage data, significantly reducing database CPU usage and enhancing system performance. The update also brings a new UI feature that shows total usage analytics, providing clearer insights into data utilization.
In addition to infrastructure improvements, LiteLLM has expanded its support for a wide range of new and existing models. Notable among these are the newly supported models from Vertex AI, such as gemini-2.0-flash-lite, and Google AI Studio, alongside support for image generation and transcription capabilities. These updates aim at bolstering the flexibility and capability of LiteLLM for diverse Artificial Intelligence applications.
