Are the FLUX 3 Image weights open?

FLUX 3 Image weights are not openly downloadable yet. Black Forest Labs sells commercial weights to enterprises for self-hosted deployment and fine-tuning, and reports at launch said a public open-weight version will follow in the coming weeks. Until then, most people use FLUX 3 Image through the BFL API, the BFL playground or partner platforms such as fal.

What resolution can FLUX 3 Image generate?

FLUX 3 Image renders natively up to 5456 × 3072 pixels, about 16.8 megapixels, according to Black Forest Labs' model page. That covers 4K output in a single pass, so fine detail such as small text or fabric texture does not need a separate upscaling model afterwards.

How do bounding boxes work in FLUX 3 Image?

FLUX 3 Image reads layout as boxes on a normalized 0–1000 grid, written as [y_min, x_min, y_max, x_max]. The BFL docs show edit requests with a source box and a target box, so an object can be moved, replaced or recolored. Because the grid is normalized, the same layout description works at any output size.

How is FLUX 3 Image different from the July FLUX 3 announcement?

Black Forest Labs announced FLUX 3 in July 2026 as one multimodal model family for video, images, audio and robot actions. FLUX 3 Video went public in August and FLUX 3 Action followed in September. FLUX 3 Image, released October 1, is the image generation and editing product, adding bounding-box layout, ten reference images and native 4K output.

Black Forest Labs · 2026-10-01 · major

FLUX 3 Image — Black Forest Labs lays out images with bounding boxes

FLUX 3 Image is Black Forest Labs' image generation and editing model. You place each element with a bounding box, mix up to 10 reference images, and render native 4K. It is out now on the BFL API, with open weights promised in the coming weeks.

FLUX 3 Image launch banner from Black Forest Labs

Draw where things go, and FLUX 3 Image paints the picture to match.

Quick facts

Maker	Black Forest Labs
Released	2026-10-01
Max output	5456 × 3072 px (16.8 MP)
Reference images	Up to 10
Layout control	Bounding boxes on a 0–1000 grid
Availability	BFL API, playground, fal; enterprise weights

What is it?

FLUX 3 Image brings layout control to Black Forest Labs' FLUX 3 family: you tell the model where each element goes with a bounding box, then it generates or edits the image to match. It is the image part of the multimodal FLUX 3 model that BFL first announced in July, and it launched on October 1, 2026.

How does it work?

Boxes are written as [y_min, x_min, y_max, x_max] on a 0–1000 grid, so the same layout works at any resolution. For edits, a source box and a target box let the model recolor, replace or move one element while the rest of the picture stays the same. Up to ten reference images can feed one generation, and BFL says the model is built so an LLM can write the layout from a short prompt.

Why does it matter?

Designers and product teams usually fight image models over composition. Explicit boxes and region-locked edits make results repeatable across many editing steps, and native 4K output removes a separate upscaling pass for print or large screens.

Who is it for?

designers, ad and e-commerce teams, developers building image tools

Frequently asked questions

Are the FLUX 3 Image weights open?: FLUX 3 Image weights are not openly downloadable yet. Black Forest Labs sells commercial weights to enterprises for self-hosted deployment and fine-tuning, and reports at launch said a public open-weight version will follow in the coming weeks. Until then, most people use FLUX 3 Image through the BFL API, the BFL playground or partner platforms such as fal.
What resolution can FLUX 3 Image generate?: FLUX 3 Image renders natively up to 5456 × 3072 pixels, about 16.8 megapixels, according to Black Forest Labs' model page. That covers 4K output in a single pass, so fine detail such as small text or fabric texture does not need a separate upscaling model afterwards.
How do bounding boxes work in FLUX 3 Image?: FLUX 3 Image reads layout as boxes on a normalized 0–1000 grid, written as [y_min, x_min, y_max, x_max]. The BFL docs show edit requests with a source box and a target box, so an object can be moved, replaced or recolored. Because the grid is normalized, the same layout description works at any output size.
How is FLUX 3 Image different from the July FLUX 3 announcement?: Black Forest Labs announced FLUX 3 in July 2026 as one multimodal model family for video, images, audio and robot actions. FLUX 3 Video went public in August and FLUX 3 Action followed in September. FLUX 3 Image, released October 1, is the image generation and editing product, adding bounding-box layout, ten reference images and native 4K output.

Try it

https://fal.ai/models/bfl/flux-3/image