Kimi K2.5

Kimi K2.5 is Moonshot AI’s top open-source reasoning model with SOTA performance and 2× faster INT4 inference.

Top open-source reasoning model, delivering state-of-the-art performance while running faster in INT4 inference compared with earlier versions. It is optimized for long-horizon tasks, complex workflows, and robust tool use, making it a powerful engine for coding agents and analytical assistants. The model delivers strong benchmark results while keeping runtime and token usage within practical production limits. It is well-suited for organizations seeking frontier-like capabilities in an open model.

image of technology-driven data analysis scene

Explore other AI Models