Google DeepMind has launched Genie 4, a world model capable of generating photorealistic interactive environments in real time, complete with sound and speaking characters.
The research preview became available to trusted testers on October 27, according to a blog post by the AI research laboratory.
Users can control simulations through text prompts as they unfold, adjusting weather conditions, changing camera angles or introducing people and objects into the environment.
DeepMind said Genie 4 was its first real-time world model capable of generating audio natively, including music, ambient sounds and dialogue that responds to users’ actions.
“Understanding and simulating the world is a crucial aspect of building AGI,” the DeepMind team wrote.
Genie 4 builds on Google’s Gemini models, with Gemini Omni responsible for how virtual worlds look and behave, while Gemini 3.8 Flash acts as an agent that monitors the environment, breaks tasks into steps and checks each new frame to confirm progress.
In one experiment, Gemini 4 Argon, which Google released publicly this month, planned a task for Genie 4 to simulate: Apptronik’s Apollo humanoid robot playing checkers. DeepMind said the technology could create an extensive range of training scenarios for robots.
Genie 4 can maintain consistent environments for up to two minutes, twice the duration supported by Genie 3, according to TechCrunch’s Rebecca Bellan.
The laboratory also said the model offers more realistic physics, including aerodynamics and collisions, while enabling more natural human movement.
According to DeepMind, testers across film, journalism, education, gaming and engineering have used Genie 4 for previsualisation and game prototyping. Some photojournalists have also experimented with converting archival photographs into interactive simulations.
The model will initially be available to Google AI Ultra subscribers through Project Genie, while selected early partners will gain access through the Gemini API and Google AI Studio. DeepMind has opened a waitlist but has yet to announce a launch date.
Google AI Ultra subscriptions start at $100 per month, according to Ars Technica’s Ryan Whitwam. The publication also noted that the unveiling of Genie 3 last year triggered a decline in gaming stocks.
