Moonshot releases Kimi K3.1 Pro with long context and faster variant
Moonshot AI introduces Kimi K3.1 Pro, a 3T-class open-weight model with a one-million-token context window and a faster Highspeed variant.
#
Key takeaways
- Moonshot AI says Kimi K3.1 Pro is its first K3.1-series model.
- The company describes it as a 3T-class open-weight model with a one-million-token context window.
- Moonshot says a Highspeed variant is two to three times faster.
- Both versions are available in web and mobile products, Kimi Work, Kimi Code, and the API.
- Moonshot plans to publish the full model weights by November 4.
Moonshot launches Kimi K3.1 Pro
Moonshot AI has released Kimi K3.1 Pro, according to its official Kimi research listing dated October 24, 2026. The company describes it as the first model in the K3.1 series and a post-trained successor to Kimi K3. It also calls the model a 3T-class open-weight system.
The source says Kimi K3.1 Pro includes a one-million-token context window. Moonshot says the model improves coding, 3D creation, financial research, and long-running agent workflows. These are company claims from the official technical page, so they should be read as Moonshot's own reported positioning.
Highspeed variant and product access
Moonshot also announced a Highspeed variant. The company says this version is two to three times faster and is designed for more responsive interaction. The source does not provide independent verification of that speed claim, so it remains a vendor-reported comparison.
Moonshot says both versions are currently available in its web and mobile products, Kimi Work, Kimi Code, and its API. The source also says the company plans to publish the full model weights by November 4. That future release should be treated as planned, not completed.
Reported evaluations and safety notes
Moonshot reports a 65.7 percent score on Terminal-Bench 4.0. It also says its own cyber-risk-reduction training lowers performance on its internal exploit benchmark when compared with a checkpoint that does not use those measures. The source frames both results as company-reported evaluations with specific setups.
That matters for interpretation. The reported score does not establish general performance across all tasks. The internal benchmark result also does not prove safety in every setting. It only shows Moonshot's own measured trade-off under the stated evaluation setup.
What the source does and does not say
The official source distinguishes current product access from the future downloadable weights. It confirms that users can access the model through Moonshot's products and API now. It also confirms that the full weights are planned for later publication.
The source does not state any Moroccan availability. It also does not mention a local commercial partnership or a Morocco-specific benchmark. For readers, the main lesson is conditional and global: when a model launch includes both product access and a later weight release, it is important to separate what is available now from what is only planned.
Morocco relevance
The source reports no Morocco-specific fact. If readers are evaluating similar launches, the useful global lesson is to check whether access, weights, and benchmark claims refer to the same release stage.
Bottom line
Kimi K3.1 Pro is presented by Moonshot as a major update in its K3.1 line. The company emphasizes long context, stronger task performance, and a faster variant. It also provides specific benchmark claims and a clear timeline for the future weight release, all under Moonshot's own reporting.
Follow us on Google
Add Intelligence Artificielle Maroc as a preferred source to see more of our relevant stories in Google Search.
What would you like to build?
We build custom AI platforms, SaaS products, intelligent business applications, and automation systems.
This form is for project inquiries, not general questions about artificial intelligence.
