Kimi 3.1 Pro
An open-weight large language model released by Moonshot AI that uses a sparse (MoE) architecture, emphasizing a large parameter pool with fewer activated parameters to control inference cost, and accompanied by a high-speed API.
Timeline (last 90 days)
Doubao Work has opened third-party models to enterprises, initially supporting Kimi 3.1 Pro, GLM 5.3, and DeepSeek v4.1 Flash.
The Kimi 3.1 Pro specifications circulating in the community show that it uses a 104BA-2.8T architecture and is an open-weight model.
According to community reports, Kimi 3.1 Pro narrows the gap between open-weight models and top closed-source models to about one month.
Kimi 3.1 Pro comes with a Pro Highspeed API, claimed to be 2-3 times faster.
All Facts (4)
The Kimi 3.1 Pro specifications circulating in the community show that it uses a 104BA-2.8T architecture and is an open-weight model.
50%UnverifiedAccording to community reports, Kimi 3.1 Pro narrows the gap between open-weight models and top closed-source models to about one month.
50%UnverifiedKimi 3.1 Pro comes with a Pro Highspeed API, claimed to be 2-3 times faster.
50%UnverifiedDoubao Work has opened third-party models to enterprises, initially supporting Kimi 3.1 Pro, GLM 5.3, and DeepSeek v4.1 Flash.
50%