Moonshot AI has released the Kimi K3.1 Pro model with a 1 million token context window

Moonshot AI has released the weights for its flagship multimodal model, Kimi K3.1 Pro. The new model is available to developers on the Hugging Face platform.

The model's architecture is based on the MoE (Mixture of Experts) approach and features 2.8 trillion parameters. A key feature of the system is an expanded context window of 1 million tokens, which allows for the processing of massive text documents and multimodal data in a single request.

The release is aimed at practitioners and researchers working on resource-intensive long-context analysis tasks. You can access the model and review the technical specifications directly via the project repository.