FLUX 3 Multimodal Foundation Model Controls Factory Robots

thione · x · 2026-07-27

Black Forest Labs announced early access for FLUX 3, its new multimodal foundation model jointly trained on images, video, and audio. Video prediction accounts for over 95% of compute costs.

Additionally, BFL partnered with mimic to develop FLUX-mimic. This video-action model leverages FLUX 3's physical world understanding to control factory robots via demonstration learning, already tested and deployed at Audi.

Related event: FLUX 3 Multimodal Model Crosses Over to Robot Control(2 posts)→

Original post →

More from Embodied

Embodied channel →