Google DeepMind unveils Genie 4 world model with real-time audio and scene editing
Google DeepMind announced Genie 4, a world model available in research preview that generates interactive environments from prompts. It adds real-time audio, more realistic worlds and the ability to steer or edit scenes with natural-language prompts, including changes during a simulation. Genie 4 can keep a scene consistent for up to two minutes, double the limit reported for Genie 3, and improves physics and character realism. Google is pitching it for creative prototyping and robot-training scenarios.
