AdvertiseDonate
Wednesday, November 18, 2026
Big-city news, all in one place.
Hot News Report · 835 Hanover Street, Suite 201, Manchester, NH 03104

Moonshot AI cuts model costs by 33% as Kimi K3.1 Pro scores 55

The new reasoning model ranks fourth globally, closing the performance gap with top competitors to just three points while reducing API fees.

By Yuki Melnyk ·

Illustration for: Moonshot AI cuts model costs by 33% as Kimi K3.1 Pro scores 55
Illustration generated for this article, not a news photograph.

Developers using Moonshot AI’s new Kimi K3.1 Pro model will pay less to run their applications starting today. The company reduced API input prices to $2 per million tokens and output prices to $10 per million tokens, a 33 percent decrease from previous rates. Cache hit costs dropped to $0.20 per million tokens. These price cuts apply to both the standard K3.1 Pro and the K3.1 Pro Highspeed variant, which Moonshot says runs two to three times faster.

Kimi K3.1 Pro now ranks fourth on the Artificial Analysis Intelligence Index with a score of 55 points. This represents an 11 point improvement over its predecessor, Kimi K3. Only three models currently score higher on the index: Anthropic’s Claude Opus 5.5 at 58 points, Claude Sonnet 5.5 at 56 points, and Muse Wonder Full at 56 points. The new model is a post trained version of K3 that utilized substantially more post training compute.

Moonshot plans to publish the model weights by November 4, 2026. If released as scheduled, K3.1 Pro would become the highest scoring open weights model available. This would narrow the gap between open and proprietary models from 12 points to 3 points, assuming no other changes occur in the interim. The model architecture remains unchanged from K3, featuring 2.8 trillion total parameters with 104 billion active parameters and a one million token context window.

Performance gains were most visible in agentic tasks. On the GDPval AA benchmark, K3.1 Pro reached an Elo score of 1,846, which is 312 points above K3. It trails only Claude Opus 5.5 at 1,866 and Muse Wonder Full at 1,857 on that specific test. On Terminal Bench 4.0, Moonshot reports a score of 65.7 percent. This result surpasses GPT 6 Astra at 58.2 percent and sits slightly below Claude Opus 5.5 at 66.4 percent. The previous K3 model managed only 13 percent on that benchmark.

The company also reported a sharp decline in hallucinations. On its internal Kimi Hallu Bench Hard, claim level hallucinations fell from 23.4 percent in K3 to 8.3 percent in the new version. Moonshot describes this benchmark as a stress test rather than a measure of general error rates. GPT 6 Astra currently leads that specific internal test at 5.5 percent. In finance, the new model gained 7.2 points on Financial Modeling and 11.6 points on Financial Research benchmarks.

K3.1 Pro includes new capabilities for 3D modeling and coding. The model works with game engines such as Unity, Unreal Engine, and Godot. It can write and run scripts for Blender and CAD software without needing MCP connectors. On the BenchCAD benchmark, it scores 92.7 percent with tools enabled. Without tools, it reaches 72.8 percent, compared to 73.0 percent for Claude Opus 5.5. Moonshot added a 3D Make feature on Kimi.com to allow users to preview, edit, and export 3D models.

Moonshot addressed cybersecurity concerns in its technical blog post. The company removed cyber related data from both training phases for the open weights release. 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 on the same test. Moonshot stated that it wants to raise the capabilities of open models gradually to support defensive work like auditing legacy systems, noting that no open weights model should push the absolute frontier of offensive cyber capabilities.

Both versions of the model are available through Kimi.ai, Kimi Work, Kimi Code, and the Kimi API. The standard version and the Highspeed variant share the same core architecture. The cost of running each Intelligence Index task is $1.98, nearly identical to the $2.00 cost for K3, because the new model generates 14 percent more output tokens.

More Tech · Report a business · Got a Story?

Get Hot News Report in your inboxBig-city news, all in one place. Free, a few times a week, unsubscribe any time.

Related stories

Get Hot News Report in your inboxBig-city news, all in one place. Free, a few times a week, unsubscribe any time.