FLUX 3 Image for precise layouts

A conventional prompt can describe a poster, but exact placement of its title, product and background objects can be difficult to control. FLUX 3 Image adds a layout method: place each important element in a bounding box on a normalized canvas and describe what belongs there. Black Forest Labs’ model page says people can draw these boxes in the Playground or pass a layout prompt to the API. Boxes are optional, so ordinary text-to-image remains available for briefs that do not need a rigid composition.

The same product supports image-to-image work and targeted editing. A team can use a reference, change one selected region or rearrange parts of a scene without intentionally rebuilding everything else. This makes it relevant to advertising layouts, editorial graphics and product variations where a consistent frame matters. The model page shows examples of typography and composed scenes, but a published example is not proof that every requested word, product shape or brand element will survive a revision. Outputs still require visual inspection, especially for rights-sensitive likenesses, logos and exact copy.

BFL charges per completed image, with rates varying by selected resolution rather than a subscription seat. The launch’s 50% introductory price ended October 8, 2026; the official list-price schedule begins at $0.041 for 768-by-768 generation and rises to $0.607 at 4K. Reference images and prompts were described as included at no extra charge in the pricing notes. API and Playground use the model under BFL terms; teams seeking high volume or private weights should discuss enterprise arrangements rather than assume FLUX.2 self-serve licenses automatically cover FLUX 3 Image. Keep this image product separate from the wider FLUX 3 video and action research programs.