# Kimi K3.1 Pro and Kimi K3.1 Pro Highspeed are available now in AI Gateway

Source: [Vercel](https://vercel.com/changelog/kimi-k3-1-pro-and-kimi-k3-1-pro-highspeed-are-available-now-in-ai-gateway)  
Feed7 permalink: https://feed7.dev/p/kimi-k3-1-pro-and-kimi-k3-1-pro-highspeed-are-available-now-in-ai-gatewa-1gne8bl  
Published: 2026-10-23T00:00:00.000Z  
Trust: Official Source (official_source)

## Why Included

Kimi K3.1 Pro and its faster Highspeed variant are available through AI Gateway for coding-agent use, with image input and a 1M-token context window.

## Source Summary

Moonshot AI’s **Kimi K3.1 Pro** and **Kimi K3.1 Pro Highspeed** are now available through Vercel AI Gateway. Both retain image input and a **1M-token context window**; Highspeed is positioned for faster interactive use.

## Practical Implication

Builders can select either model in Claude Code, Codex, Cursor, or other agents configured through the latest Vercel CLI. The standard model targets coding, long multi-step agent runs, 3D work, and financial analysis.

## Agent-Ready Context

Moonshot AI’s **Kimi K3.1 Pro** and **Kimi K3.1 Pro Highspeed** are now available through Vercel AI Gateway. Both retain image input and a **1M-token context window**; Highspeed is positioned for faster interactive use.

Builders can select either model in Claude Code, Codex, Cursor, or other agents configured through the latest Vercel CLI. The standard model targets coding, long multi-step agent runs, 3D work, and financial analysis.

The material provides no latency, price, or coding benchmark results for comparing the variants. Moonshot stated that model weights were planned for release by **November 4**, so the availability described here is initially gateway-based.

## Connected Context

Feed7 judgment across 898 accumulated Signals:

This expands the crowded million-token agent-model pool with two image-capable variants separated mainly by intended interaction speed. Because no latency, price, or quality results compare them, the Highspeed label does not yet establish a routing rule. The planned weight release also leaves gateway access as the initially supported path rather than confirming a local deployment option.

- [Hy4 Preview now available on AI Gateway](https://feed7.dev/p/hy4-preview-now-available-on-ai-gateway-0knvicj) — Both add 1M-context coding-agent routes through the same gateway, reinforcing the need to compare sustained-context quality, latency, cost, and reliability on representative tasks.
- [Muse Spark 1.3 now available on AI Gateway](https://feed7.dev/p/muse-spark-1-3-now-available-on-ai-gateway-0b1bnu3) — Muse Spark 1.3 is another 1M-token multimodal agent option, so Kimi’s image support and context capacity alone do not narrow selection without matched evaluation.
- [unslothai/unsloth](https://feed7.dev/p/unsloth-1l373r0) — If Kimi’s planned weights become available in a compatible form, local serving tools such as Unsloth could provide an alternative to gateway-only access; the supplied material does not confirm that deployment path yet.

## Context Map

- Layer: model
- Domains: coding, data
- Topics: model-selection, coding-agents, open-models

## Uncertainty

- The material provides no latency, price, or coding benchmark results for comparing the variants. Moonshot stated that model weights were planned for release by **November 4**, so the availability described here is initially gateway-based.

## Agent Instruction

Use this item as source-backed context. Do not invent claims beyond the linked source. If this item conflicts with another source, call out the conflict.