Google unveils Genie 4 AI model for generating interactive worlds

Summarised from 3 outlets · Updated 27 Oct, 22:45 · Archive

Google DeepMind released Genie 4, a world model that generates more realistic interactive experiences with audio and real-time editing.

Google DeepMind has released Genie 4, an updated world model that generates interactive virtual experiences. The model creates worlds frame by frame in real time as users explore them, without requiring pre-rendered environments or pre-programmed interactive elements.

Genie 4 has several significant improvements over its predecessor, Genie 3. It can generate more photorealistic worlds with better physics simulation and more human-like characters. Users can edit scenes mid-simulation by typing natural language prompts to change elements such as weather or camera angle. The model also generates real-time audio that matches the simulated world, including music, ambient noises and interactive dialogue. Experiences can now last up to two minutes, double the length of Genie 3's one-minute limit.

The model uses Gemini 3.8 Flash, Google's AI model, as a guiding agent that breaks down goals and evaluates each output frame to keep the simulation on track. This integration helps maintain context as users move through varying environments.

Google is pitching Genie 4 for applications including pre-visualization for filmmakers, prototyping game ideas, and creating training scenarios for robots. The company notes that as simulations run longer, some details and object permanence can shift over time. Simulations remain consistent for about two minutes before losing track of how things started.

Genie 4 is launching first as a 'research preview' to 'trusted testers'. The company has opened a waitlist, and the model will be coming soon to Google AI Ultra subscribers.

Google unveils Genie 4 AI model for generating interactive worlds — Image: Ars Technica

Comments (0)

How it is being reported

Google’s Genie 4 world model adds audio, more realistic worlds, and real-time steeringGoogle DeepMind released Genie 4, a world model with real-time audio, more realistic simulations, and mid-simulation text prompts, in research preview for trusted testers.TechCrunch · 27 Oct, 16:30
DeepMind unveils Genie 4 world model promising more realistic physics and humansGenie 4 will roll out to developers and Google AI Ultra subscribers soon.Ars Technica · 27 Oct, 16:00
Google’s next Genie AI model generates more realistic experiencesGoogle announced today its Genie 4 world model for generating interactive experiences using AI. The company says the model has four main updates: It allows for more realistic worlds, lets users update an experience in real time with a natural language prompt, can generate audio, and uses Gemini 3.8 Flash under the hood to help […]The Verge · 27 Oct, 16:00

In this story: Google DeepMind · Gemini

This summary was written by AI from the headlines and standfirsts above, and states as fact only what at least two outlets report. How we use AI · Report a problem · Corrections log