What happened
When Moonshot AI launched Kimi K3.1 Pro in October, it said the weights would follow. The moonshotai/Kimi-K3.1-Pro repository on Hugging Face was created on November 2, 2026, according to the Hugging Face API; that is a creation date, and the exact moment the files became public is not stated. The model card calls it "an open-weight, natively multimodal agentic model".
What is in the release
The card lists a mixture-of-experts model with 2.8 trillion total and 104 billion activated parameters, 896 experts with 16 selected per token, text and image input and a context length of 1,048,576 tokens. Weights use MXFP4 with MXFP8 activations, trained with quantization in mind. Moonshot recommends vLLM, SGLang or TokenSpeed for serving; on its own API the model is kimi-k3.1-pro, with reasoning effort low, high or max, max by default.
Moonshot's own results at max effort include 65.7 on Terminal-Bench 4.0, 94.2 on GPQA-Diamond, 92.1 on BrowseComp and 70.0 on HLE-Full with tools. These are the company's figures, with its own harness choices, not independent tests.
The license terms
The Kimi K3.1-Pro License lets you use, copy, modify, distribute and fine-tune the model, with conditions. If you or your affiliates run a Model as a Service business and your combined revenue exceeds $20 million over any 12 months, you must sign a separate agreement with Moonshot AI before any commercial use. A product with more than 100 million monthly active users or more than $20 million in monthly revenue must prominently display "Kimi K3.1-Pro". Neither condition applies to internal use or to access through Moonshot's own products or certified inference partners.
What it means for you
Small teams and researchers can run K3.1 Pro on their own hardware without a separate deal, though a model this size needs serious multi-GPU infrastructure. If you resell model access as an API and are above the revenue threshold, contact Moonshot before launch. Otherwise the Kimi API remains the simpler route.
Go to the source
Sources & further reading
Based on the official documentation below. Features and access can change; check the provider before making a decision.
Was this page useful?
Help us make our coverage more useful. This is feedback on the page, not a rating of the tool.
Go a little further