Chinese AI model Kimi K3.1 Pro is closing in on Anthropic's Claude Opus 5.5
decoder ·
Chinese AI company Moonshot AI has released Kimi K3.1 Pro, a new reasoning model that ranks fourth on the Artificial Analysis Intelligence Index with 55 points. This score places it just three points behind Anthropic's Claude Opus 5.5, marking an 11-point improvement over its predecessor, Kimi K3. The model demonstrates significant gains in agentic tasks, coding, and 3D modeling while maintaining a parameter count of 2.8 trillion. Token prices have dropped by 33 percent, making the API more cost-effective without sacrificing performance. Moonshot also introduced a faster variant called K3.1 Pro Highspeed for users requiring quicker response times. In a move to address cybersecurity concerns, the company removed specific cyber-related data from the training set before releasing the open weights. This strategic reduction aims to balance offensive capabilities with defensive utility for auditing legacy systems. If released as planned in November 2026, K3.1 Pro would become the highest-scoring open-weights model available.
decoder