Vercel adds Black Forest Labs' FLUX 3 Image to AI Gateway

October 5, 2026Source: Vercel

Vercel has made FLUX 3 Image from Black Forest Labs available on its AI Gateway. Developers can generate images from text or edit existing ones with the same model, using a single API key and no Black Forest Labs account.

Key points

FLUX 3 Image from Black Forest Labs is now available on Vercel's AI Gateway.
One AI Gateway API key is enough; no Black Forest Labs account is required.
The model supports text-to-image and editing, with up to ten reference images.
Five resolution tiers run from 768 × 768 to 4K, in square, portrait and landscape ratios.
It works with AI SDK 7 or later using the model ID bfl/flux-3-image, and with AI CLI.

What FLUX 3 Image offers on AI Gateway

According to Vercel, FLUX 3 Image handles both text-to-image generation and image editing with the same model. Access runs through AI Gateway, so developers do not need to register separately with Black Forest Labs.

For editing, the model supports up to ten reference images. Vercel says these can be used for editing, combining subjects, or changing the composition of an image.

Output comes in five resolution tiers, ranging from 768 × 768 to 4K. Square, portrait and landscape aspect ratios are supported.

How to use it in code

Setup requires the AI_GATEWAY_API_KEY environment variable and the model ID bfl/flux-3-image. The model works in AI SDK 7 or later.

To generate an image, developers call generateImage from the ai package with a text prompt. Vercel's example asks for a blue ceramic vase with wildflowers on a sunlit wooden table. It sets a 1:1 aspect ratio and passes a Black Forest Labs provider option of resolution: '1k' to produce a square, 1K-resolution image.

Editing uses the same function. Instead of a plain string, the prompt becomes an object with a text instruction and an images array. In the changelog example, the generated vase image is passed back in with the instruction to change the vase to terracotta while keeping the flowers and setting.

Terminal access with AI CLI

FLUX 3 Image can also be used from the command line with AI CLI, using the same AI Gateway API key and model ID. Vercel's instructions are short:

Install the tool with npm install -g ai-cli.
Run ai image with a prompt, set --model bfl/flux-3-image, and use --output to choose where the result is saved, for example ./renders/.

Vercel also points to a model playground where FLUX 3 Image can be tried, a page with model details, and an image generation guide.

What it means for developers

The change is mainly about access. Teams already working with AI Gateway and the AI SDK can add a Black Forest Labs image model by switching a model ID, rather than setting up a separate account and credentials. Because generation and editing share one model and one call pattern, an application can produce an image and then revise it without changing providers.

The source does not include pricing or rate limit details, so developers will need to check the model details page for those.