Moonshot AI's Kimi K3.1 Pro Ranks Fourth Globally, Closing Gap with Claude Opus 5.5 at Half the Cost

Moonshot AI has released Kimi K3.1 Pro, a new 3-trillion-parameter language model that ranks fourth among 215 models on the Artificial Analysis Intelligence Index, trailing only Claude Opus 5.5 in overall capability while offering significantly cheaper pricing. Announced on October 23, 2026, the model features a 1-million-token context window, multimodal input support for text and images, and an aggressive pricing structure at $2 per million input tokens and $10 per million output tokens—undercuts Anthropic's Opus 5.5 by half on input costs and one-third on output costs. Moonshot trained K3.1 Pro as a substantially enhanced version of its predecessor K3, with expanded post-training compute across hundreds of specialized domains to address coding, finance, and 3D modeling tasks.