Google has announced Genie 4, a world model designed to create interactive simulations.
The model builds on capabilities from Google’s Gemini family. Gemini Omni helps simulate real-world appearances and behavior, while other Gemini models support reasoning to maintain consistency across simulated scenes.
With Genie 4, users can change scenes in real time using natural language commands, such as moving the camera, changing the weather, or adding characters.
The model can also generate audio, including ambient sounds, music, and interactive dialogue.
Genie 4 is currently rolling out to a limited group of trusted testers.
Google plans to make it available to Google AI Ultra subscribers through Project Genie, along with early access partners through the Gemini API and AI Studio.