Moonshot AI's New Model Cuts Prices and Narrows Gap With Top Rivals
Kimi K3.1 Pro boosts performance in coding and 3D tasks while reducing API costs by one-third, positioning it just behind the industry leaders.
By Jian Chen ·
Moonshot AI has officially launched Kimi K3.1 Pro, a new reasoning model that immediately secured a top-five spot on the Artificial Analysis Intelligence Index. The release marks a significant shift in the competitive landscape, as the model scores 55 points, placing it just three points behind the current leaders. This jump of 11 points over its predecessor, Kimi K3, signals a rapid acceleration in the company's post-training capabilities and positions the Chinese firm as a serious contender in the global AI market.
The performance gains are most visible in complex, agentic tasks. On the GDPval-AA benchmark, the new model achieves an Elo rating of 1,846, a substantial 312-point increase from the previous version. It now trails only Claude Opus 5.5 and Muse Wonder Full in this specific metric. Furthermore, the model demonstrates strong coding abilities, recording a 65.7 percent success rate on Terminal-Bench 4.0. This figure surpasses GPT-6 Astra and sits nearly on par with Claude Opus 5.5, whereas the older K3 model managed only 13 percent on the same test.
Beyond raw coding power, Moonshot has significantly reduced the model's tendency to hallucinate. On its internal Kimi Hallu Bench Hard, claim-level hallucinations dropped from 23.4 percent in K3 to 8.3 percent in K3.1 Pro. The company notes that this internal stress test is not a general measure of error rates but highlights a clear improvement in reliability. While GPT-6 Astra still holds an edge with a 5.5 percent rate, the reduction suggests the model is becoming more precise in its factual outputs for enterprise applications.
A major differentiator for K3.1 Pro is its capability in 3D modeling and direct integration with game engines. The model can interact with Unity, Unreal Engine, and Godot to handle everything from minor fixes to full project builds. It scores 92.7 percent on the BenchCAD benchmark when using tools, approaching the performance of top proprietary models. Users can also access a new feature called "3D Make" on Kimi.com, which allows them to preview, edit, and export 3D models directly, streamlining workflows for developers and designers.
Despite the performance boost, the model maintains the same architectural scale as its predecessor. It retains 2.8 trillion total parameters, with 104 billion active parameters, and supports a one-million token context window for both text and image inputs. According to Artificial Analysis, this places the model at the top of the Pareto frontier, meaning no other model of comparable size currently scores higher on the Intelligence Index. This efficiency is a critical factor for developers looking to balance cost with capability.
Pricing has also been adjusted to reflect the improved efficiency. API costs have dropped by 33 percent across the board. Input tokens are now priced at $2 per million, down from $3, while output tokens cost $10, reduced from $15. Cache hits are cheaper at $0.20 per million tokens. Although the model generates 14 percent more output tokens per task, the overall cost per Intelligence Index task remains nearly identical to the previous version at $1.98, ensuring that the performance upgrade does not come with a significant price tag for developers.
In preparation for an open-weights release planned for November 4, 2026, Moonshot has taken deliberate steps regarding the model's cybersecurity capabilities. The company removed cyber-related data from both training phases to manage offensive potential. On an internal benchmark that reproduces recent vulnerabilities in software like the Linux kernel, the uncleaned version solved 90.5 percent of tasks. After data cleaning, K3.1 Pro solves 81.0 percent, a figure that exceeds the older K3 but remains below DeepSeek-V4.1-Flash. This approach aims to balance offensive capability with defensive utility, such as auditing legacy systems.
The launch of Kimi K3.1 Pro and its faster variant, K3.1 Pro Highspeed, is available now through Kimi.ai, Kimi Work, Kimi Code, and the Kimi API. If the planned release of model weights proceeds as scheduled, K3.1 Pro would become the highest-scoring open-weights model available. This move would shrink the performance gap between open and proprietary models from 12 points to just three, fundamentally altering the trade-off between cost, access, and capability for businesses and developers in the region.
Source: The Decoder
