The listed model IDs are moonshotai/kimi-k3.1-pro and moonshotai/kimi-k3.1-pro-highspeed.
Vercel says K3.1 Pro is the first model in Moonshot's K3.1 series. It retains a one-million-token context window and image input, and the company claims improvements in coding, 3D modeling, financial analysis, and agentic tasks. It says the model can be called through AI SDK, Chat Completions, Responses, and Anthropic Messages APIs, or selected through coding agents connected to AI Gateway.
For agent setup, Vercel instructs users of Claude Code, Codex, Cursor, and similar tools to update Vercel CLI and run npx vercel ai-gateway setup. The company says Gateway passes through provider pricing without markup or an inference platform fee, including BYOK requests.
Moonshot plans to release the model weights by November 4. The announcement does not provide benchmark scores, token prices, availability outside the named Gateway endpoints, or independent evidence for the stated capability gains.