Google DeepMind has introduced Genie 4, a world model capable of generating interactive simulated environments that users can steer in real time through natural-language instructions. The model creates visuals and native audio together, including environmental sounds, music and interactive dialogue.
Users can alter a generated environment while it is running by changing the weather, moving the camera, introducing characters or setting events in motion. DeepMind said Genie 4 incorporates capabilities from the Gemini family: Gemini Omni helps anchor simulations in real-world appearance and behaviour, while other Gemini models support reasoning, continuity and scene coherence.
The technology is positioned for applications extending beyond entertainment. DeepMind identified robotics and AI-agent training, filmmaking, gaming, journalism and manufacturing as potential areas where interactive simulations could be used. For robotics and autonomous systems, generated worlds could provide controlled training environments without exposing physical equipment or people to early-stage system failures.
Genie 4 remains a limited research release rather than a generally available commercial product. Access has started with a small group of trusted testers. Google plans to make it available through Project Genie for Google AI Ultra subscribers, with early access also planned for selected Gemini API and Google AI Studio partners. No India-specific availability schedule has been announced.
