Moonshot Introduces Kimi K3.1 Pro Model
Moonshot has introduced the Kimi K3.1 Pro model, which is expected to be open weight by November 4, 2026, and offers competitive pricing and performance. The model has shown promising results in benchmarks, scoring highly against other models.
Original reporting · simonwillison.net
Original: “Introducing Kimi K3.1 Pro”
Read “Introducing Kimi K3.1 Pro” at Simon Willison →
How this was written. Content Memory wrote this article with a language model from Simon Willison’s reporting (model: @cf/meta/llama-3.3-70b-instruct-fp8-fast); it describes that reporting rather than adding to it. The original, linked above, is the full story.
Key points from Simon Willison’s report
- Moonshot has introduced the Kimi K3.1 Pro model, which is expected to be open weight by November 4, 2026
- The model offers competitive pricing and performance, scoring highly in benchmarks against other models
- The Kimi K3.1 Pro model is available on the Kimi API, with a 'Highspeed' version offering 2-3x faster performance at twice the price
- The model's performance and pricing make it an attractive option for developers and organizations looking to utilize AI capabilities
- The release of the model's full weights is eagerly anticipated, as it will allow developers to access and utilize the model's capabilities more freely
According to Simon Willison, Moonshot has introduced the Kimi K3.1 Pro model, which is now available on the Kimi API. The model is not yet open weight, but Moonshot promises to release the full model weights by November 4, 2026. This move is significant as it will allow developers to access and utilize the model's capabilities more freely.
The Kimi K3.1 Pro model offers competitive pricing, with US dollar pricing set at $2.00/million tokens for input and $10/million for output. This pricing is the same as GPT-6.1 Sol and half the price of Claude Opus 5.5. Additionally, Moonshot is offering a 'Highspeed' version of the model, which is 2-3x faster, but at twice the price, except for input token cache hits, which are priced the same.
Moonshot's own benchmarks show that the Kimi K3.1 Pro model performs approximately level with Opus 5.5 and GPT-6 Astra, beating them on a few benchmarks and coming close on others. This suggests that the Kimi K3.1 Pro model is a strong contender in the field. Artificial Analysis has also scored the Kimi K3.1 Pro model at 55, ahead of Claude Fable 5.1, GPT-6 Astra, and Gemini 4 Argon, and behind only Claude Opus 5.5.
The introduction of the Kimi K3.1 Pro model is notable as it is the first time a soon-to-be open weight model has scored this highly against the frontier. This could have significant implications for the development of AI models and their applications. Simon Willison reported that he tried out the Kimi K3.1 Pro model via the Moonshot API using the llm openai endpoint command and was able to generate an SVG of a pelican riding a bicycle.
The ability to run prompts against any OpenAI-compatible endpoint using the llm openai endpoint command is a useful feature, allowing developers to test and utilize different models. However, the command does not log the results, requiring additional options to save stdout and stderr. The Kimi K3.1 Pro model's performance and pricing make it an attractive option for developers and organizations looking to utilize AI capabilities.
The release of the Kimi K3.1 Pro model's full weights by November 4, 2026, is eagerly anticipated, as it will allow developers to access and utilize the model's capabilities more freely. This could lead to further innovation and development in the field of AI, as well as increased adoption of the Kimi K3.1 Pro model.
The Kimi K3.1 Pro model's competitive pricing and performance, combined with its soon-to-be open weight status, make it a significant development in the field of AI. As the model becomes more widely available, it will be interesting to see how it is utilized and what applications it enables.
Questions and answers
When will the full model weights of Kimi K3.1 Pro be released?
The full model weights of Kimi K3.1 Pro are expected to be released by November 4, 2026.
How does the pricing of Kimi K3.1 Pro compare to other models?
The pricing of Kimi K3.1 Pro is the same as GPT-6.1 Sol and half the price of Claude Opus 5.5.
What is the 'Highspeed' version of Kimi K3.1 Pro?
The 'Highspeed' version of Kimi K3.1 Pro is 2-3x faster than the standard version, but at twice the price, except for input token cache hits, which are priced the same.
