Google DeepMind announced Genie 4, a world model that generates interactive environments and accompanying audio. The preview is rolling out to trusted testers, rather than opening immediately to every consumer or developer.
Sound joins the simulated world
DeepMind says the model generates ambient noise, music and dialogue alongside visual scenes. Users navigate with keyboard controls and steer scenes through text prompts.
Access remains staged
Google says broader access is coming to Google AI Ultra subscribers through Project Genie and early-access partners using the Gemini API and Google AI Studio. A waitlist is available, but the announcement does not establish general availability.
Consistency is still a limitation
DeepMind acknowledges drifting visual details and object permanence over longer sessions or complex edits. It also identifies physical interactions, accurate audio, legible text and spatial consistency as unfinished research areas. Generated video and audio will carry SynthID watermarks, according to Google. The demonstrations show research progress, not proof that the system can reliably replace a production game engine or validated physical simulator.
Sources
Google DeepMind: Genie 4 research announcement
Feature image: AI-generated editorial illustration for Guava Insights. Conceptual artwork, not an actual product photograph, interface or facility.
