Kimi K3.1 Pro launches via API, with open weights promised for November 4
The weights were not released at launch, so only the API is available for now.
2.8TTotal parameters in the new Kimi model
The short version
- Moonshot launched Kimi K3.1 Pro, a multimodal mixture-of-experts model with 2.8 trillion total parameters, 104 billion active parameters, and a 1-million-token context window.
- Standard and Highspeed versions launched on Kimi's website and app, Kimi Work, Kimi Code, and the API. Moonshot says Highspeed generates output 2–3× faster.
- At launch the weights were not published. Moonshot's announcement promises them by November 4, 2026.
- Artificial Analysis placed K3.1 Pro fourth on its Intelligence Index with a score of 55, behind Opus 5.5 at 58. Standard first-party pricing is $2 per million input tokens and $10 per million output tokens, down from K3's $3 and $15.
Why it matters
If the weights are released as promised, Artificial Analysis says the gap between leading open and proprietary models would narrow from 12 points to three. For developers, the model is a large one: the article gives an illustrative estimate of about 1.4 TB for the parameters alone at 4-bit, before serving overhead. The article notes that self-hostable and consumer-device-local are different properties.
“we optimized every layer of the stack”
What people are saying
Commenters on Hacker News questioned the removal of cyber-related training data from the open-weight release. Some pointed to a drop in hallucination on one benchmark, from 23.4% for K3 to 8.3% for K3.1, and called the jump impressive. One commenter asked whether the launch was why the AI trade was weak that week.