A soon-to-be open-weights model now scores near the frontier at about half the price of Opus 5.5 per token. Engineers choosing models or planning self-hosting can weigh it, with weights promised by Nov 4, 2026.
Key takeaways · AI-distilled
Simon Willison reports Kimi K3.1 Pro at $2/$10 per million input/output tokenThe chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.Full definition →, matching GPT-6.1 Sol. The Highspeed variant is 2-3x faster at twice the price, except cache-hit input tokens, which cost the same.
Per Willison, Artificial Analysis scored Kimi K3.1 Pro at 55, ahead of Claude Fable 5.1, GPT-6 Astra and Gemini 4 Argon (all 53). He calls it the first time a soon-to-be open-weights model has scored this high against the frontier.
Moonshot's own benchmarks put K3.1 Pro roughly level with Opus 5.5 and GPT-6 Astra, beating them on a few benchmarks and coming close on others.
Willison tested it with LLMA large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.Full definition →'s llm openai endpoint command, which works with any OpenAI-compatible API but does not log results, so he saved stdout and stderr (the reasoning trace) to files. The model offers low, medium and max reasoning levels.
Terms in this piece · Glossary
token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
LLM — A large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.
