Kimi K3.1 Pro and Highspeed variant arrive on Vercel's AI Gateway
Vercel has added Moonshot AI's Kimi K3.1 Pro to its AI Gateway, together with Kimi K3.1 Pro Highspeed, a faster version meant for interactive use. K3.1 Pro is the first model in Moonshot's K3.1 series and builds on Kimi K3.
- Vercel added Moonshot AI's Kimi K3.1 Pro and Kimi K3.1 Pro Highspeed to AI Gateway.
- K3.1 Pro is the first model in the K3.1 series and improves on Kimi K3 in coding, 3D modeling, financial analysis and agentic tasks.
- The model keeps the 1M-token context window and image input.
- Moonshot AI plans to release the model weights by November 4, with more providers to follow.
- AI Gateway reflects provider pricing with no markup and charges no platform fee on inference.
What K3.1 Pro improves
According to Vercel, Kimi K3.1 Pro improves on Kimi K3 in coding, 3D modeling, financial analysis and agentic tasks. It keeps the 1M-token context window and image input of its predecessor.
For coding, the model covers work ranging from small fixes in an existing codebase to building a full project, including long agent runs with many steps. Vercel says it builds front-end pages and interactive scenes with stronger typography and animation than earlier Kimi models.
In finance, the model can carry a task from start to finish. Vercel gives two examples: building a linked financial model with adjustable assumptions and scenarios, and writing research grounded in source documents. The post also states that K3.1 Pro makes fewer factual errors than earlier Kimi models on difficult prompts.
Weights and access
Moonshot AI plans to release the model weights by November 4, according to the post. Support for more providers will follow.
Developers can call either model through the AI SDK, the Chat Completions, Responses and Anthropic Messages APIs, or a coding agent connected to AI Gateway. The model name for the standard version is moonshotai/kimi-k3.1-pro. For the faster variant, developers use moonshotai/kimi-k3.1-pro-highspeed instead.
Vercel's example uses the AI SDK's streamText function with the prompt "Write a Blender script that models a parametric bookshelf." The model can also be tried in the model playground.
Using it in coding agents
To use the models in Claude Code, Codex, Cursor and similar tools, Vercel tells users to install the latest Vercel CLI and run npx vercel ai-gateway setup. They then select either moonshotai/kimi-k3.1-pro or moonshotai/kimi-k3.1-pro-highspeed in the agent. A coding agents guide covers the details.
What AI Gateway offers
AI Gateway is Vercel's unified API for calling models. According to the company, it tracks usage and cost and lets users configure retries, failover and performance optimizations for higher-than-provider uptime. Other listed features include:
- built-in custom reporting
- Zero Data Retention support
- budgets for API keys
- routing rules
Vercel says AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests.
For developers already routing traffic through AI Gateway, the release means K3.1 Pro can be tested by changing the model name, with a faster option available for interactive applications. The planned weight release suggests broader provider support is coming, though the source gives no further details.
