Google releases Genie 4 world model with audio and real-time steering

According to TechCrunch, Genie 4 can maintain a consistent scene for up to about two minutes, double Genie 3’s one-minute limit, and generates real-time audio that matches the simulated world, including music, ambient noises and interactive dialogue. Users can type prompts during a simulation to change the weather or camera angle or add people and objects. Google says the model offers better physics and more human-like characters.

The Verge reported that Genie 4 uses Gemini 3.8 Flash under the hood. TechCrunch reported that Gemini acts as a director that plans steps and checks frames while Genie creates the visuals. In a research experiment tied to Gemini Robotics, a Gemini 4 Argon model broke tasks into steps that Genie 4 simulated in real time, showing a possible way to test robot scenarios before they happen in the real world. Tejas Kulkarni, research director at Google, said during a briefing that future systems will increasingly be hybrid world models, combining Gemini writing code and text with Genie ingesting that in real time and simulating reality at different levels of abstraction.

Google is pitching Genie 4 for pre-visualization by filmmakers, prototyping game ideas, training scenarios for robots and interactive learning. Diego Rivas, a Google DeepMind engineer, said during a press briefing that the real-time aspect is most valuable in creative work where keeping a state of flow matters, such as ideation and storyboarding. He said users could eventually rely on Gemini Omni to turn prototypes into creative assets, but those capabilities are still a ways off.

Google warned that as simulations run longer, details and object permanence can shift over time. TechCrunch reported that the team acknowledged objects can drift and that it has not fully solved how to ensure audio accuracy or legible in-world text.

Genie 4 is available now as a research preview for trusted testers, according to TechCrunch. Google says the model will come soon to Google AI Ultra subscribers in Project Genie, its tool for generating interactive worlds from text or image prompts, and has opened a waitlist. TechCrunch reported that it will also roll out in the coming weeks to early access partners of the Gemini API and Google AI Studio.

The release comes as competition in world models intensifies. TechCrunch reported that AI video startup Runway released its second world model, GWM Worlds 2, a couple of months earlier; that AMD is acquiring Fei-Fei Li’s World Labs; that Amazon is in talks to buy Decart; and that Rivian has acquired Third Dimension to bring simulation capabilities in-house for self-driving testing.