Genie 4 Unveiled for Real-Time Simulation

Google DeepMind just dropped Genie 4, a world model that lets you steer a live simulation with a single voice command. It plugs into Gemini Omni to anchor scenes in real-world look and feel, while Gemini models keep the story coherent. You can change the weather, move the camera, or drop new characters on the fly, all in natural language. Genie 4 even generates native audio-ambient sounds, music, and interactive dialogue-without pre-recorded clips. This tech is a step toward AGI, giving robots and agents a ground boosting creativity in film, gaming, journalism, and manufacturing. Google is rolling it out to a small group of trusted testers now. Soon it will hit Google AI Ultra subscribers in Project Genie and early-access partners in Gemini API and AI Studio. Check the demos on the DeepMind blog to see the model in action.