Black Forest Labs Releases FLUX 3 Image with 4K Resolution and Bounding Box Control
Black Forest Labs has announced FLUX 3 Image, an advanced image generation model supporting up to 4K resolution output and precise spatial layout using bounding box coordinates. The model allows users to fuse up to 10 reference images in a single prompt and perform pixel-precise regional edits. Spatial layout control has historically been a major challenge for generative text-to-image models. By introducing explicit coordinate positioning and reference-image fusion, FLUX 3 Image offers the granular control needed for professional graphic design, advertising, and commercial workflows. Users can arrange visual elements on a normalized 0–1000 coordinate grid by providing bounding box coordinates formatted as [y_min, x_min, y_max, x_max] alongside prompts and IDs for each element. The model is currently offered as a paid service, with open-weight versions planned for release within a few weeks.
## BACKGROUND
Black Forest Labs is an AI lab founded by key researchers behind Stable Diffusion, known for developing state-of-the-art visual generation models under the FLUX brand. Bounding box control in generative AI allows users to specify exactly where objects should appear on the canvas, solving issues where text prompts fail to capture precise layout requirements.