Moonshot AI releases Kimi K3.1 Pro weights for its 2.8-trillion-parameter model Moonshot AI, led by co-founder and chief executive Yang Zhilin, published the weights for its 2.8-trillion-parameter Kimi K3.1 Pro model on Hugging Face on November 4th, 2026, under a Kimi-specific license, fulfilling a promise the Beijing company made when it introduced the model in October. The release lets developers download and self-host the model rather than relying only on Moonshot AI's API, while the Kimi-specific license keeps limits on that openness. Moonshot AI announced the release in a post on X linking to the Hugging Face repository, adding: "We'll keep advancing open intelligence. Moonshot AI releases Kimi K3.1 Pro weights for its 2.8-trillion-parameter model Yang Zhilin's Moonshot AI published the 2.8-trillion-parameter model on Hugging Face on November 4th, under a Kimi-specific license. By Ryan Merket https://runtimewire.com/author/ryan-merket · Published · Updated Primary source: Kimi / Moonshot AI official https://x.com/Kimi Moonshot/status/2118010528123773297 Why it matters The release lets developers download and serve Moonshot AI's most capable Kimi model instead of relying only on its API, while the Kimi-specific license keeps limits on that openness. Outside teams can now test Moonshot AI's coding and agent claims under their own harnesses, and customers gain a self-hosting option to weigh against hosted pricing. Moonshot AI, led by co-founder and chief executive Yang Zhilin, released the model weights for Kimi K3.1 Pro https://runtimewire.com/models/moonshotai/kimi-k3.1-pro on November 4th, 2026, completing a promise the Beijing company made when it introduced the model in October. Moonshot AI announced the release in a post on X https://x.com/Kimi Moonshot/status/2118010528123773297?ref=runtimewire that linked to the Kimi K3.1 Pro repository on Hugging Face https://huggingface.co/moonshotai/Kimi-K3.1-Pro?ref=runtimewire and added: "We'll keep advancing open intelligence." The model card describes Kimi K3.1 Pro as… When you purchase through links in our articles, we may earn a small commission https://runtimewire.com/affiliate-monetization-standards . This doesn’t affect our editorial independence.