According to Wccftech, Moonshot is planning to launch the Kimi K3.1 model, which is expected to debut next month (October 2026) and will offer three levels of reasoning intensity: Low, High, and Max. This information comes from a source named @MaxForAI on the X platform, who discovered it through internal JSON/API responses from Moonshot — including model names such as "k3d1-agent," and snippets also revealed Agent mode, application scenarios, and various configuration switches.

Exposed Configuration: Extended Context, Multi-Agent, and Swarm

According to the relevant configuration files, K3.1 may support multiple capabilities: multiple levels of reasoning intensity such as Low, High, and Max; an "Extra Long" option for extended context, supporting up to 1 million tokens; Agent mode; Swarm multi-agent collaboration; and task modes such as search and batch processing. If these configurations are accurate, Kimi K3.1 will further enhance its capabilities in extended context, agent collaboration, and task scheduling.

Overview of the Previous Flagship Model K3's Parameters and Architecture

As a reference, the flagship model Kimi K3 currently released by Moonshot was launched in July 2026. Official data shows that it has 2.8 trillion parameters, native visual capabilities, and a 1 million token context window, and it uses architectural designs such as Kimi Delta Attention, Attention Residuals, and sparse MoE. On the basis of K3, what specific metrics will K3.1 upgrade, and whether it can continue to maintain the "flagship" positioning, remains to be revealed after Moonshot officially releases it.