| Mobile Web

Search results for multimodal AI

AI & Enterprise

Black Forest Labs unveils Flux 3, eyes robotics beyond video generation

Black Forest Labs has unveiled Flux 3, its first video-generation AI model that can create video and audio at the same time. The multimodal model can generate up to 20 seconds of video along with dialogue, sound effects and background music. The company said Flux 3 extends beyond content creation toward “Physical AI” and robot control via Flux-mimic, developed with Zurich-based Mimic Robotics. Audi is testing the technology on a production line.

AI & Enterprise

Nvidia joins Japan\'s FRONTIA national physical AI project

Japan\'s government is working with Nvidia to build national artificial intelligence infrastructure aimed at strengthening competitiveness in physical AI used in robotics and industrial automation. Up to 1 trillion yen is expected to be invested over the next five years. The economy ministry will support cooperation between consortium NOESTRA and Nvidia on the FRONTIA project, which runs until 2030 and will use Nvidia\'s Vera Rubin platform infrastructure.

AI & Enterprise

Google Veo adds personal AI avatars from selfies and voice

Google has added a personal AI avatar feature to its video creation tool Veo, allowing users to generate customised digital avatars that reflect their face and voice. Users can upload a selfie and a voice recording to insert the avatar into videos. Google also applied its multimodal AI model Gemini Omni to Veo, supporting text prompts and reference images, background replacement and lighting correction. Editing now allows step-by-step changes without recreating a video.