Google DeepMind Unveils Genie 4, a Real-Time World Model
Google DeepMind calls Genie 4 its most realistic and interactive world model yet. You explore dynamic environments in real time: the scene responds, sounds play, and people move, react and speak. You can steer a scene live with natural language, such as moving the camera, changing the weather or adding new characters, and every world makes native audio, including dialogue and music. Output carries a SynthID watermark. It is a research preview with trusted testers now, and it is coming soon to Google AI Ultra subscribers in Project Genie (waitlist open) and to early access partners in the Gemini API and AI Studio. Google has not announced pricing; Ars Technica notes Google AI Ultra starts at $100 a month. Money angle, ideas only and no results yet: explorable ad scenes for brands, teaser trailers, previews of game worlds before a team builds them, and B-roll you steer instead of search for.
Join the Project Genie waitlist
Watch the DeepMind video →Google blog post →Gemini tool page →Veo tool page →
Source: Google DeepMind (official, X and blog) and Google (official blog)
