Kimi Releases K3.1 Pro Model Weights

What happened

Kimi announces the release of Kimi K3.1 Pro's model weights and links to the model repository on Hugging Face. The post adds, “We'll keep advancing open intelligence.”

Read next

• More changes in today’s briefingOpenAI Releases Images 2.6 With SVG Export, GIFs, and Faster Generation

The story so far

Moonshot AI released Kimi K3.1 Pro’s model weights on Hugging Face, following its earlier launch through Kimi’s apps and API. The weights make the model openly available; Artificial Analysis’s revised Intelligence Index still ranks it ninth, behind eight closed models. K3.1 Pro initially ranked fourth with a score of 55, three points behind Claude Opus 5.5. In version 5, Artificial Analysis says it scores 46 and trails the leader by 12 points. The revision assigns 60% of the index to private test sets. Ethan Mollick urged caution about the wider gap reported between open and closed models. On Artificial Analysis’s separate Cyber Index, K3.1 Pro scored 64%, leading openly available models for agentic cyber defense, while trusted-access GPT-6 Sol (Daybreak Blue, max) scored 69%. Artificial Analysis reported no safety refusals across the three benchmarks. Its cost per task was $2.81, 15% higher than Kimi K3 despite 33% lower per-token prices.