The Pulse

Kimi K3.1 Pro adds more ZDR providers on Vercel AI Gateway

Kimi K3.1 Pro and K3.1 Pro Highspeed are now available on Vercel AI Gateway from more providers with Zero Data Retention agreements. The page describes their capabilities, ways to call them and the Gateway settings that require ZDR for requ

AI.info Team · Written with AI · fact-checked · 2026-11-06

Kimi K3.1 Pro and its higher-speed version, Kimi K3.1 Pro Highspeed, are now available on Vercel AI Gateway from more Zero Data Retention (ZDR) providers. Until this change, Moonshot AI was the only provider serving either model.

K3.1 Pro improves on earlier Kimi models in coding, 3D modeling and financial analysis, and can handle long agent runs, from a small fix to a full project. Both K3.1 Pro and Highspeed have a 1M-token context window and accept image input. Vercel describes Highspeed as suited to interactive work such as coding agents, where response time matters.

Calling the models

Both models can be called through the AI SDK, the Chat Completions, Responses and Anthropic Messages APIs, or a coding agent connected to AI Gateway. The model name for K3.1 Pro is moonshotai/kimi-k3.1-pro. For Highspeed, use moonshotai/kimi-k3.1-pro-highspeed in place of that model name.

Vercel’s example uses the AI SDK’s streamText function and sets the model name and prompt in the request. The prompt asks for a Python function that redacts email addresses from a log line. The example also sets zeroDataRetention: true in the gateway provider options:

import { streamText } from 'ai';

const result = streamText({ model: 'moonshotai/kimi-k3.1-pro', prompt: 'Write a Python function that redacts email addresses from a log line.', providerOptions: { gateway: { zeroDataRetention: true, }, },
});

For coding agents, the announcement lists Claude Code, Codex and Cursor among the supported options. Users can install the latest Vercel CLI and run vercel ai-gateway coding-agents setup, then select either moonshotai/kimi-k3.1-pro or moonshotai/kimi-k3.1-pro-highspeed in the agent. The CLI installation command shown is pnpm i -g vercel@latest. Vercel also directs users to try Kimi K3.1 Pro in the model playground and points to its coding agents guide for setup details.

How ZDR routing works

Setting zeroDataRetention: true in the AI Gateway provider options requires ZDR for that request. Teams can instead turn on ZDR in the AI Gateway dashboard settings to require it for every request they send, including requests from coding agents. With either setting enabled, AI Gateway routes only to providers that have a ZDR agreement. If no such provider is available for a model, the request fails rather than being sent to a provider without an agreement.

Vercel describes AI Gateway as a unified API for calling models and tracking usage and cost. It also offers configuration for retries, failover and performance optimizations intended to provide higher-than-provider uptime, as well as custom reporting, ZDR support, API-key budgets and routing rules. The service reflects provider pricing without a markup and does not charge a platform fee on inference, including for Bring Your Own Key requests.