Moonshot's New AI Model Narrows Gap With Top Rivals

Moonshot AI has launched Kimi K3.1 Pro, a new reasoning model that has rapidly climbed the global rankings to challenge the leading proprietary systems. The release represents a significant step forward for the Chinese company, which aims to close the performance gap with established Western competitors. By significantly increasing the amount of compute used during the post-training phase, the new model demonstrates substantial improvements over its predecessor, Kimi K3.

The model currently sits at number four on the Artificial Analysis Intelligence Index, scoring 55 points. This marks an 11-point jump from the previous version. It trails only Claude Opus 5.5, Claude Sonnet 5.5, and Muse Wonder Full, which hold the top three spots. If Moonshot follows through on its plan to release the model's weights by November 4, 2026, it will become the highest-scoring open-weights model available. This move would shrink the performance distance between open and proprietary models from 12 points to just three, assuming no other major updates occur in the interim.

The most notable gains for K3.1 Pro appear in agentic tasks, where the model can execute complex, multi-step workflows. On the GDPval-AA benchmark, the new model reaches an Elo score of 1,846, a 312-point increase over K3. This places it just behind Claude Opus 5.5 and Muse Wonder Full. In coding environments, specifically Terminal-Bench 4.0, Moonshot reports a success rate of 65.7 percent. This figure surpasses GPT-6 Astra and nearly matches Claude Opus 5.5, whereas the older K3 model managed only 13 percent on the same test.

Reliability has also improved, with a sharp drop in hallucination rates. According to Moonshot's internal stress test, claim-level hallucinations fell from 23.4 percent in the previous version to 8.3 percent. While this is a significant improvement, GPT-6 Astra still holds an edge with a lower rate of 5.5 percent on the same benchmark. The company notes that this specific test is a stress test rather than a measure of general error rates, but it indicates a more grounded approach to factuality.

The model also shows strong capabilities in three-dimensional modeling and finance. It can interact directly with game engines like Unity, Unreal Engine, and Godot for both quick fixes and full project builds. On the BenchCAD benchmark, it scores 92.7 percent when using tools, a result close to the top proprietary models. Moonshot states that the model is particularly strong at parametric modeling for mechanical parts and architectural designs. Additionally, it can write and run scripts for software like Blender without requiring external connectors.

Despite the performance boost, the model retains the same architectural size as its predecessor. It maintains 2.8 trillion total parameters, with 104 billion active ones, and supports a context window of one million tokens. However, the cost to use the model has decreased by 33 percent. Input tokens now cost $2 per million, down from $3, while output tokens are priced at $10 per million, reduced from $15. This makes the high-performance model more accessible for developers and businesses.

Moonshot has also taken specific steps to manage the model's cybersecurity capabilities. The company observed that vulnerability research skills grew as post-training scaled, so they removed cyber-related data from the training process for the open-weights release. This adjustment reduced its ability to solve specific security tasks on an internal benchmark from 90.5 percent to 81.0 percent. While this is still high, the company aims to balance offensive capabilities with defensive uses, such as auditing legacy systems, while ensuring no single open model dominates the frontier of offensive cyber capabilities.

A faster variant, called K3.1 Pro Highspeed, is also available now. It runs two to three times faster than the standard version. Both models are accessible through various platforms, including Kimi.ai, Kimi Work, and the Kimi API. The release signals an intensifying competition in the AI sector, where performance gains are increasingly paired with cost reductions and strategic decisions about how much capability to make publicly available.