Moonshot Adds Kimi K3.1 Pro to API

Moonshot added Kimi K3.1 Pro to its API on October 23, 2026. The model is not yet available with open weights, but Moonshot says it will release the full weights by November 4, 2026. Moonshot also offers a “Highspeed” version that it says runs 2–3x faster and costs twice as much, except cached input tokens remain the same price. Standard pricing is $2.00 per million input tokens and $10 per million output tokens, matching GPT-6.1 Sol and half the price of Claude Opus 5.5.

Moonshot’s benchmarks place Kimi K3.1 Pro approximately level with Opus 5.5 and GPT-6 Astra, beating them on some tests and coming close on others. Artificial Analysis gave it a score of 55, ahead of Claude Fable 5.1, GPT-6 Astra and Gemini 4 Argon, each at 53, and behind Claude Opus 5.5 at 58. The author called this the first time a soon-to-be open-weight model had scored this highly against frontier models.

Simon Willison tested Kimi K3.1 Pro through Moonshot’s API using the `llm openai endpoint` command added to LLM 0.32 in August. The command can send prompts to any OpenAI-compatible endpoint, but does not log results; Willison used extra options to save standard output and the reasoning trace. He generated an SVG of a pelican riding a bicycle at the model’s three supported reasoning levels—low, medium and max—and compared those results with fresh generations from Kimi K3 at the same three levels.

Moonshot added Kimi K3.1 Pro to its API on October 23, 2026. The model is not yet available with open weights, but Moonshot says it will release the full weights by November 4, 2026. Moonshot also offers a “Highspeed” version that it says runs 2–3x faster and costs twice as much, except cached input tokens remain the same price. Standard pricing is $2.00 per million input tokens and $10 per million output tokens, matching GPT-6.1 Sol and half the price of Claude Opus 5.5.

Moonshot’s benchmarks place Kimi K3.1 Pro approximately level with Opus 5.5 and GPT-6 Astra, beating them on some tests and coming close on others. Artificial Analysis gave it a score of 55, ahead of Claude Fable 5.1, GPT-6 Astra and Gemini 4 Argon, each at 53, and behind Claude Opus 5.5 at 58. The author called this the first time a soon-to-be open-weight model had scored this highly against frontier models.

Simon Willison tested Kimi K3.1 Pro through Moonshot’s API using the `llm openai endpoint` command added to LLM 0.32 in August. The command can send prompts to any OpenAI-compatible endpoint, but does not log results; Willison used extra options to save standard output and the reasoning trace. He generated an SVG of a pelican riding a bicycle at the model’s three supported reasoning levels—low, medium and max—and compared those results with fresh generations from Kimi K3 at the same three levels.