Google's Genie 4 Adds Native Audio to Real-Time World Models

Google has introduced Genie 4, a world model that creates interactive environments in real time and generates native audio, including ambient sound, music and dialogue. Users can steer a scene with natural-language instructions, changing the weather, moving the camera or adding people and objects while the simulation continues.

The system combines several models from Google's Gemini family. Gemini Omni anchors appearance and behavior, other Gemini models help maintain coherence, and Gemini 3.8 Flash can act as a guide that observes a scene, breaks a goal into steps and tracks progress. Google is also experimenting with Genie 4 and another Gemini model for robotics simulations.

The launch is still a research preview, not a generally available commercial product. Google has begun offering it to a limited group of trusted testers in creative industries, journalism, education, gaming and engineering. Broader access is planned for Google AI Ultra subscribers through Project Genie and for early partners using the Gemini API and Google AI Studio, but the company did not announce pricing or a general-release date.

Google also identified important limits. Longer sessions can drift visually, and complex physical interactions, audio accuracy, text rendering and long-term consistency remain difficult. Generated video and audio carry SynthID watermarking, giving users a way to identify AI-produced material.

For Alphabet investors, Genie 4 widens the company's AI platform from text, images and video toward simulated environments that could support creative tools, games and robot training. The investment case is still indirect. The preview demonstrates capability and potential distribution through paid products and developer platforms, but it provides no usage, cost, revenue or customer-conversion data.