Wire · founder news, decoded · opportunities
Black Forest Labs Unveils FLUX 3, A New Multimodal Frontier Model For Visual Intelligence
◆ Published
23 July 2026
◆ Topic
opportunities
◆ Geography
◆ Source
◆ Verified
Fusion42 · 23 July 2026 · Fusion42 review
Black Forest Labs unveils FLUX 3, a multimodal frontier model jointly trained on images, video, audio, and action prediction in a unified architecture. The model is already being tested by creative platforms (Canva, Picsart, Krea) and manufacturing companies (Audi via mimic robotics), with early evaluations showing performance leadership in video generation and robotic manipulation tasks.
◆ The Wire takeaway
If you're building robotic perception or video generation, your foundation model just got cheaper to train on. FLUX 3 cuts robot training data from 30 hours to 30 minutes per task—that's your manufacturing sales cycle accelerated by 60x.
◆ Related on Wire
- Glass crashes slashed? Ant Group embodied AI unit claims breakthrough in robot sensing7 July 2026
- Ant Group Subsidiary Robbyant Releases First MoE Video Foundation Model for Robotics9 July 2026
- BMW expands humanoid robotics software at Landshut plant21 July 2026
- Mistral launches first robotics model in physical AI push9 July 2026
- China Visual AI Leader SenseTime Declares End of Stitch-Monster Era With Open-Source ...17 July 2026
- China's Moonshot unveils new AI model17 July 2026
◆ Topics
foundation-models · multimodal-learning · robotics-ai · video-generation · action-prediction