Chinese AI Model Threatens Anthropic's Lead in Global Intelligence Rankings

Chinese AI firm Moonshot AI has released Kimi K3.1 Pro, a model that now ranks fourth on the Artificial Analysis Intelligence Index with a score of 55. The release places the model just three points behind Anthropic's Claude Opus 5.5, signaling a rapid narrowing of the performance gap between proprietary systems and those developed by Chinese competitors. The model represents a significant jump over its predecessor, Kimi K3, rising 11 points on the index.

The ranking places K3.1 Pro directly behind Claude Opus 5.5, Claude Sonnet 5.5, and Muse Wonder Full. Moonshot AI has scheduled the release of K3.1 Pro model weights for November 4, 2026. If this release proceeds as planned, K3.1 Pro would become the highest-scoring open-weights model available, reducing the gap between open and proprietary models from 12 points to just 3. Alongside the standard release, Moonshot is also shipping a variant called K3.1 Pro Highspeed, which the company claims runs two to three times faster than the standard model.

The model's capabilities in agentic tasks, coding, and 3D modeling have seen substantial improvements. On the GDPval-AA benchmark, K3.1 Pro achieved an Elo of 1,846, a 312-point increase over K3, placing it just behind Claude Opus 5.5 at 1,866. On the Terminal-Bench 4.0 test, Moonshot reports a score of 65.7 percent, compared to 66.4 percent for Claude Opus 5.5, while GPT-6 Astra scored 58.2 percent. The model also demonstrated a significant reduction in hallucination rates; on Moonshot's internal Kimi Hallu Bench Hard, claim-level hallucinations dropped from 23.4 percent in K3 to 8.3 percent.

In specific applications, K3.1 Pro shows notable capabilities in financial and creative work. On internal finance benchmarks, the model gained 7.2 points on Financial Modeling and 11.6 points on Financial Research. For general office tasks, Moonshot claims the model tops all tested models on OfficeQA Pro v2. Regarding 3D modeling, the model can operate directly with game engines such as Unity, Unreal Engine, and Godot to handle both minor fixes and full project builds. On the BenchCAD benchmark, it scores 92.7 percent with tools, close to Claude Opus 5.5's 96.2 percent.

Moonshot AI has adjusted the pricing structure for K3.1 Pro, reducing token prices by 33 percent. Input costs are now set at $2 per million tokens, down from $3, while output costs are $10, down from $15. The model maintains the same dimensions as its predecessor, with 2.8 trillion total parameters and 104 billion active ones, alongside a one million token context window.

Security concerns have prompted Moonshot AI to modify the model's training data. For the open-weights release, the company stripped cyber-related data from both training phases to limit offensive capabilities. On its internal xdev-bench-hard, 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, while DeepSeek-V4.1-Flash reaches 85.7 percent on the same test.

Moonshot AI has not confirmed if it will release further updates to K3.1 Pro before the scheduled release of model weights in November. The company states that while no open-weights model should push the absolute frontier of offensive cyber capabilities, it wants to gradually raise what open models can do for defensive purposes like auditing legacy systems. Anthropic has previously indicated that even GLM-5.3, which scores 45 on the AA Index, can write exploits at near-Claude Mythos levels.