From the source
Google DeepMind has previewed Genie 4, a real-time world model that builds photorealistic environments with fluid motion and generates ambient sound, music-like audio and spoken dialogue.
- DeepMind says Genie 4 models aerodynamics, collisions, natural human movement and reactions to surroundings, with steerable characters, and calls it the company's first real-time world model with native audio.
- Users navigate with keyboard controls and reshape a running scene through text, so a prompt can change weather, move the camera, or add a person or object without restarting.
- For multi-step scenarios, Gemini 3.8 Flash acts as a guiding agent that watches frames, divides a goal into steps, tracks progress, checks outcomes and replans when the world changes or a user intervenes.
- The research preview is rolling out to trusted testers in creative industries, journalism, education, gaming and engineering, with experiments covering film previsualization, simulations from archival 2D photos, and game concepts.
- DeepMind says generated video and audio carry SynthID watermarks, and warns that visual details and object permanence can drift during long sessions, repeated complex edits or navigation.