Google unveils Genie 4 world model for AI-generated interactive worlds

Google announced on 27 October 2026 its Genie 4 world model, an AI system that generates interactive experiences. Google DeepMind describes it as a new generation of world simulation with a higher level of realism in real-time simulation.

According to Google, the model brings four main updates. It produces more realistic worlds, lets users change an experience as it runs by typing a natural language prompt, can generate audio, and uses Gemini 3.8 Flash under the hood to help with its simulations.

The length of experiences has also increased. Genie 4 can create experiences of about two minutes, compared with one minute for the earlier Genie 3 model.

Google is presenting the model as a tool for several kinds of work. The company names pre-visualization for filmmakers, prototyping of game ideas, and the making of training scenarios for robots as possible uses.

Google also points to a limitation. It warns that as simulations run for longer, details and object permanence can shift over time.

Genie 4 is launching first as a research preview for what Google calls trusted testers. The Verge's reporter said he had not been able to try it himself, so he could not compare it with Project Genie, which is powered by Genie 3 and which he had used before.

Google says Genie 4 will be coming soon to Google AI Ultra subscribers. The company has also opened a waitlist for people who want to apply to try the model in Project Genie.

Google has published examples demonstrating what the new model can do, and DeepMind has released a blog post describing Genie 4.