Kimi K3.1 Pro lands on AI Gateway with its 1M-token window intact, plus a Highspeed variant for interactive use

Kimi K3.1 Pro and its faster variant Kimi K3.1 Pro Highspeed from Moonshot AI are now available on AI Gateway, reachable through the AI SDK, the Chat Completions, Responses and Anthropic Messages APIs, or a coding agent connected to the gateway. K3.1 Pro is the first model in Moonshot's K3.1 series, builds on Kimi K3, and keeps the 1M-token context window and image input. The stated improvements cover coding, 3D modeling, financial analysis and agentic tasks; coding ranges from small fixes in an existing codebase to building a full project, including long agent runs with many steps.

For coding, K3.1 Pro builds front-end pages and interactive scenes with typography and animation that Moonshot says are stronger than earlier Kimi models; for finance, it can carry a task from start to finish, such as a linked financial model with adjustable assumptions and scenarios or research grounded in source documents. Moonshot says K3.1 Pro makes fewer factual errors than earlier Kimi models on difficult prompts. The comparison is against earlier Kimi models rather than other vendors' models, and no benchmark scores or error rates are given.

The release note covers the model's availability on AI Gateway, how to call it, and the capability directions Moonshot describes, for developers preparing to reach Kimi K3.1 Pro through the unified API or a coding agent. Moonshot says it plans to release the model weights by November 4 and that support for more providers will follow.

The improvements in coding, 3D modeling, financial analysis and agentic tasks, and the fewer factual errors on difficult prompts, are all stated relative to earlier Kimi models without quantitative results. The weight release is described as planned for November 4, and support for more providers as following later.