Moonshot AI's Kimi K3.1 Pro Lands Fourth in Global Model Rankings

Moonshot AI, a Chinese firm, has unveiled Kimi K3.1 Pro, a reasoning model that immediately claimed fourth place on the Artificial Analysis Intelligence Index. The model's 55-point score marks an 11-point leap over its predecessor, Kimi K3, and places it just three points shy of top proprietary systems. Higher-ranked models include Claude Opus 5.5 at 58 points, alongside Claude Sonnet 5.5 and Muse Wonder Full, both at 56.

K3.1 Pro is a post-trained version of Kimi K3 built with significantly more post-training compute, yet its parameter count remains unchanged at 2.8 trillion total and 104 billion active. The model handles a one-million-token context window and processes both text and images. Moonshot also released a variant called K3.1 Pro Highspeed that runs two to three times faster, according to the company. Both versions are available now through Kimi.ai, Kimi Work, Kimi Code, and the Kimi API. Token pricing dropped 33 percent across the board, with input costs falling to $2 per million tokens from $3, output to $10 from $15, and cache hits to $0.20 from $0.30.

On agentic benchmarks, the model shows sharp gains. On Terminal-Bench 4.0, Moonshot reports 65.7 percent, ahead of GPT-6 Astra at 58.2 percent and just behind Claude Opus 5.5 at 66.4 percent, while Kimi K3 managed only 13 percent. On the GDPval-AA benchmark, K3.1 Pro reaches 1,846 Elo, up 312 points from K3. Hallucination rates on Moonshot's internal Kimi Hallu Bench Hard fell from 23.4 percent to 8.3 percent. In internal finance benchmarks, the model gained 7.2 points on Financial Modeling and 11.6 points on Financial Research. For general office tasks, Moonshot says K3.1 Pro tops all tested models on OfficeQA Pro v2.

The model can interface directly with game engines such as Unity, Unreal Engine, and Godot, handling both quick fixes and full project builds, according to Moonshot. On BenchCAD, it scores 92.7 percent with tools, close to Claude Opus 5.5 at 96.2 percent. Without tools, it reaches 72.8 percent, nearly matching Claude Opus 5.5 at 73.0 percent. Moonshot says K3.1 Pro writes and runs scripts for Blender and CAD software without MCP connectors, and the company launched a 3D Make feature on Kimi.com that lets users preview, edit, and export 3D models.

For open-weights distribution, Moonshot stripped cyber-related data from both training phases, a decision it says stems from a belief that no open-weights model should push the absolute frontier of offensive cyber capabilities. On its internal xdev-bench-hard test, which reproduces recently disclosed vulnerabilities in software like the Linux kernel, Node.js, and JavaScriptCore, the uncleaned version solves 90.5 percent of tasks. After data cleaning, K3.1 Pro solves 81.0 percent, compared with DeepSeek-V4.1-Flash at 85.7 percent and Kimi K3 at 33.3 percent.

Moonshot plans to release the model weights by November 4, 2026, which would make K3.1 Pro the highest-scoring open-weights model available and shrink the gap between open and proprietary models from 12 points to just 3, assuming no other changes by then. The model sits at the top of the Pareto frontier, according to Artificial Analysis, meaning no model with a comparable parameter count scores higher on the Intelligence Index.