Black Forest Labs Launches FLUX 3: Unifying Image, Video, Audio, and Robotics

pess_r · x · 2026-07-24

Black Forest Labs officially announces FLUX 3. Jointly trained in one unified architecture, the model handles Image, Video, Audio, and Action-Prediction.

FLUX 3 Video is now available in early access. The team also demonstrated its robotics potential: in collaboration with Mimic Robotics, they built a Video-Action Model (FLUX-mimic) on top of FLUX 3, which is already being tested on complex manufacturing tasks with Audi. This marks generative visual tech rapidly expanding into physical world control.

Related event: Black Forest Labs Launches FLUX 3: A Unified Multimodal Foundation Model(28 posts)→

Original post →

More from Embodied

Embodied channel →