Black Forest Labs unveils FLUX 3, a multimodal AI model that combines images, video, audio, and action prediction for creative tools and robotics.
Black Forest Labs has introduced FLUX 3, a new multimodal AI model. The launch took place in Germany and marks a major step in artificial intelligence. FLUX 3 is designed to handle images, video, audio, and even action prediction. The company says this will help machines understand the world more fully.
The FLUX series has already been used in creative tools like Adobe Photoshop and Picsart. FLUX 3 expands those abilities by combining different types of data in one system. This means it can generate video with sound, edit images, and predict actions. The model was trained to connect vision, sound, and movement together.
Robin Rombach, co-founder and CEO, explained the importance of vision in AI. He said that images and video show structure and change, while audio adds timing and context. “Intelligence means seeing the world, predicting what happens next, and acting,” he noted. The company believes joint training makes each skill stronger.
FLUX 3 is aimed at industries ranging from media to robotics. It can create realistic video clips, match audio to events, and guide machines in physical tasks. Early results show it is good at capturing human expressions and linking sound with actions. It also supports multiple languages for global use.
The company has partnered with mimic robotics to create FLUX-mimic. This version helps robots learn new tasks quickly, sometimes in less than an hour. Audi is already testing it in its production labs. The goal is to make manufacturing more flexible and efficient.
Developers can now apply for early access to FLUX 3 Video and FLUX 3 Action. FLUX 3 Image will be released soon, and open-weight versions are planned later this year. Black Forest Labs says it wants to share the technology widely. This will allow teams to adapt the model for their own projects.
In conclusion, FLUX 3 shows how multimodal AI can connect vision, sound, and action. It opens new possibilities in creative tools, robotics, and industry. With partners already testing it, the model could shape the future of intelligent systems. More updates are expected as wider access becomes available.
Trending
