# Moonshot launches Kimi K3.1 Pro on its API, with open weights promised by Nov 4

published: 2026-10-23T20:05:30.642Z
topic: infrastructure
source: https://simonwillison.net/2026/Oct/23/kimi-k31-pro/
url: https://aiagentnews.io/stories/moonshot-launches-kimi-k3-1-pro-on-its-api-with-open-weights-promised-by-nov-4

Moonshot has made Kimi K3.1 Pro available through its API at $2 per million input tokens and $10 per million output tokens. The company says full model weights will follow by November 4, 2026, and Artificial Analysis scored the model 55, behind only Claude Opus 5.5.

Kimi K3.1 Pro is now on the Kimi API, but it is not yet released as open weights. Moonshot says full model weights will be released by Nov 4, 2026. The company also offers a Highspeed version that it describes as two to three times faster at twice the price, though input token cache hits are priced the same.

US dollar pricing is $2.00 per million input tokens and $10 per million output tokens, the same as GPT-6.1 Sol and half the price of Claude Opus 5.5. Moonshot's own benchmarks place Kimi roughly level with Opus 5.5 and GPT-6 Astra, beating them on a few tests and coming close on others.

Artificial Analysis scored Kimi at 55, ahead of Claude Fable 5.1, GPT-6 Astra and Gemini 4 Argon at 53, and behind only Claude Opus 5.5 at 58. Simon Willison tested the model through the Moonshot API using the llm openai endpoint command in LLM 0.32, running it at low, medium and max reasoning levels.

## Why it matters if you build agents

Developers can call Kimi K3.1 Pro through an OpenAI-compatible endpoint at the same input and output price as GPT-6.1 Sol, which makes it worth testing against frontier models for agent workloads. The weights are not public yet, so teams planning to self-host should wait for the promised release before depending on it. The Highspeed tier doubles the price for faster responses, which may matter for latency-sensitive agent loops.