Google DeepMind releases Genie 4 world model

On October 28, 2026, The Verge reported that Google announced Genie 4, a world model for generating interactive experiences with AI. According to the report, Google described four main updates: more realistic worlds, the ability to update an experience in real time through a natural language prompt, generated audio, and an embedded Gemini 3.8 Flash providing assistance beneath the surface. The Verge said that an experience created by the model lasts about two minutes, compared with one minute for Genie 3. At the time of that report, the model was being made available first to trusted testers as a research preview, and The Verge said it had not yet tried the system itself.

Later the same day, Google DeepMind published its own announcement of Genie 4, describing it as a new world model that unlocks unprecedented realism and improves real-time interactive simulation. Google DeepMind said Genie 4 is natively multimodal, so audio-like music, ambient noises, and interactive dialogue can be generated at the same time as the world itself. The company added that Genie 4 incorporates and builds on capabilities from the Gemini family: Gemini Omni anchors the simulations in real-world appearance and behavior, while other Gemini models provide the complex reasoning needed to help maintain coherence and continuity. With this foundation, a user sets an event in motion and Genie 4 simulates what happens next. The announcement carried the names Raia Hadsell, Diego Rivas, and Hang Qi. Google DeepMind said Genie 4 is available now as a research preview to trusted testers, and will later be offered to Google AI Ultra subscribers and early partners on the Gemini API.