FLUX 3 Image
Black Forest Labs released FLUX 3 Image, a text-to-image model featuring precise layout control via bounding boxes, JSON element tables, and pixel-perfect editing.
summary is enough The announcement introduces useful layout-control features, but technical readers can quickly grasp the capabilities without reading the full page.
summary
Black Forest Labs has launched FLUX 3 Image, featuring native 2K/4K resolution and advanced composition control through bounding boxes and JSON element tables. Users can layout scenes precisely or let AI agents plan them, while also performing targeted pixel-perfect edits. While currently available via API and commercial licenses, an open-weights version has been promised for the coming weeks. The release highlights a growing emphasis on UI steerability and structured inputs for generative image models.
key points
- FLUX 3 Image uses a two-part layout prompt consisting of a global caption and an element table containing bounding box coordinates.
- The model supports pixel-perfect editing, allowing targeted modifications to specific boxes while leaving the rest of the image untouched.
- An open-weights release of the model has been confirmed by Black Forest Labs to launch in the coming weeks.
who should read it
AI engineers and developers interested in controllable image generation APIs and structured prompt interfaces.
discussion positive
consensus
Users praise the user experience and structured layout controls, though many are waiting for the promised open-weights local release.
pushback
Some participants report mixed results with precise editing and prompt adherence on specific tests compared to competing models.
> Open Weights version of FLUX 3 Image is launching in the coming weeks.
— https://x.com/bfl_ai/status/2105734605621825738
It shares the official confirmation from Black Forest Labs that an open-weights version is launching in the coming weeks.
Generated by an LLM from the article and thread. Check the sources before relying on it.