Google DeepMind has unveiled Genie 4, an advanced world model that significantly boosts the realism of AI-generated interactive experiences. The latest iteration introduces real-time audio generation, allowing for more immersive simulations, and enables users to steer or modify these experiences mid-simulation using natural language prompts.
This new version promises more realistic physics and human-like interactions within generated worlds. Genie 4 is set to be rolled out to developers and Google AI Ultra subscribers soon, following a research preview for trusted testers. The model leverages Gemini 3.8 Flash to power its enhanced capabilities, aiming to create more dynamic and responsive virtual environments.
Google's Genie 4 AI Model Enhances Realism with Audio and Real-Time Steering
3 stories · 3 sources
Sources
- Google’s Genie 4 world model adds audio, more realistic worlds, and real-time steering — TechCrunch,
- DeepMind unveils Genie 4 world model promising more realistic physics and humans — Ars Technica,
- Google’s next Genie AI model generates more realistic experiences — The Verge,