All tracked Moonshot AI models — pricing, context windows, lifecycle state, and release history in one place.
Kimi's most versatile model. Native multimodal architecture, vision + text, thinking and non-thinking modes. 256K context.
Thinking model based on K2. General agentic and reasoning capabilities, deep reasoning tasks.
MoE model with 1T total / 32B active params. Exceptional coding and agent capabilities. 256K context.
Original Moonshot model. 128K context. Superseded by K2 series.
Predecessor → successor chains tracked for Moonshot AI models.
Kimi's most versatile model. Native multimodal architecture, vision + text, thinking and non-thinking modes. 256K context.
MoE model with 1T total / 32B active params. Exceptional coding and agent capabilities. 256K context.
Thinking model based on K2. General agentic and reasoning capabilities, deep reasoning tasks.
Original Moonshot model. 128K context. Superseded by K2 series.