Moonshot AI · AI assistants

Kimi

Updated Kimi K3.1 Pro’s full weights are now published.

An assistant and API from Moonshot AI. Kimi K3.1 Pro is the latest K3.1 release for vision, coding and long-context work, and its full open weights are now on Hugging Face.

Best for

Testing long-document and coding work against other model families. K3.1 Pro and its Highspeed variant have separate API offerings.

Worth knowing

App allowances and API pricing are separate. API prices are listed per million tokens, with a lower rate for cached input.

A task to try

Give it a long document you know well and ask for an answer with supporting passages from different sections.

Models to know

General & coding

Kimi K3.1 Pro Updated

Moonshot’s K3.1 model for vision, coding and long-context agent work. Its full weights (a 2.8-trillion-parameter mixture of experts, 104B active) are now on Hugging Face under Moonshot’s own license. The Highspeed variant is a separate API offering.

Kimi API after a successful top-up (minimum $1); open weights on Hugging Face

General & coding

Kimi K3

The earlier K3 model remains a reference for long-context, coding and vision work. K3.1 Pro is the newer release; no current v5.0 K3 score is included in this comparison.

Kimi app and API; also offered in Perplexity, GitHub Copilot and Ollama Cloud

Compare models → · See benchmark results →

Practical guides

Latest Kimi news

No stories mention Kimi in the current feeds. Browse all AI news →

Alternatives

DeepSeek

DeepSeek

Compare API cost and accuracy on your own test cases.

Anthropic

Claude

Try the same long document and compare the quality of the analysis.

Google

Gemini

Compare document tasks within a Google-centered workflow.