FLUX 3 image by Black Forest Labs is here Decide where each element goes, with multiple references in one image, and edit specific details to make it yours Now on Magnific

What will you create with FLUX 3 Image? Start here 👇

@bfl_ai Placement is the part that matters for anyone telling a story with one image. In my small dramas the whole meaning sits in where the red object is and how far apart two people stand, and that's exactly what a prompt used to scramble.

@bfl_ai Multi-reference control in FLUX 3 is a game changer

@bfl_ai The interesting test is whether those references stay coherent through several edits, especially object geometry and lighting. Would be useful to see a before/after chain with the same reference locked.

@bfl_ai Being able to decide where each element goes gives the generation process much more creative control.

@bfl_ai Number 6 overall, and number 1 for text coherence according to 6M human judgements. 👀 (PS: they'll be by far 1 open-weight if they open this same model version 🙄🙄) Believe it or not, the results are visible on congrats @bfl_ai team!

@bfl_ai