Moonshot AI Cuts Kimi K3.1 Pro Costs by a Third While Closing Performance Gap

Chinese AI developer Moonshot AI has released Kimi K3.1 Pro, a reasoning model that currently ranks fourth on the Artificial Analysis Intelligence Index. The model earned a score of 55, an 11 point increase over its predecessor, Kimi K3. This places it three points behind the leading proprietary systems.

According to Artificial Analysis, only Anthropic’s Claude Opus 5.5, Claude Sonnet 5.5, and Muse Wonder Full score higher. Moonshot plans to publish the model weights by November 4, 2026. If released on schedule, K3.1 Pro would become the highest scoring open weights model available. That move would narrow the performance gap between open and proprietary models from 12 points to 3.

Performance gains are most evident in agentic tasks. On the GDPval AA benchmark, K3.1 Pro reached 1,846 Elo, a 312 point jump from K3. It trails only Claude Opus 5.5 and Muse Wonder Full on that metric. On Terminal Bench 4.0, Moonshot reports a score of 65.7 percent. This exceeds GPT 6 Astra at 58.2 percent and sits slightly below Claude Opus 5.5 at 66.4 percent. The previous version, K3, managed only 13 percent on that test.

The company also reports a sharp drop in hallucinations. On its internal Kimi Hallu Bench Hard, claim level hallucinations fell from 23.4 percent with K3 to 8.3 percent. Moonshot describes this as a stress test rather than a general error measure, noting that GPT 6 Astra still leads at 5.5 percent. Internal finance benchmarks show gains of 7.2 points in Financial Modeling and 11.6 points in Financial Research.

K3.1 Pro works with game engines including Unity, Unreal Engine, and Godot. It can write and run scripts for Blender and CAD software without MCP connectors. On the BenchCAD benchmark, it scores 92.7 percent with tools, compared to 96.2 percent for Claude Opus 5.5. Moonshot has added a 3D Make feature on Kimi.com for previewing, editing, and exporting models.

The architecture remains unchanged from K3, featuring 2.8 trillion total parameters, 104 billion active parameters, and a one million token context window. API prices fall by 33 percent, setting input costs at $2 per million tokens and output at $10. Cache hits cost $0.20. Because the model generates 14 percent more output tokens, the cost per Intelligence Index task is $1.98, nearly identical to K3’s $2.00.

In its technical blog, Moonshot addresses cybersecurity capabilities. It states that vulnerability research skills grew as post training scaled up. For the open weights release, the company removed cyber related data from both training phases. On its internal xdev bench hard test, the uncleaned model solved 90.5 percent of tasks, while the cleaned version solved 81.0 percent. DeepSeek V4.1 Flash scored 85.7 percent and Kimi K3 scored 33.3 percent.