Google’s next Genie AI model generates more realistic experiences

Google announced today its Genie 4 earth model for generating interactive experiences using AI. The business says the example has four chief updates: It allows for additional realistic worlds, lets users update an cognition in genuine period alongside a natural tongue prompt, can create audio, and uses Gemini 3.8 Flash under the hood to assistance alongside its simulations. It additionally can create experiences that are concerning two minutes long, up from one infinitesimal for the former Genie 3 model.

Google is pitching Genie 4 as a tool that could be helpful for things akin pre-visualization for filmmakers, prototyping equivalent ideas, and making training scenarios for robots. But the business warns that, as simulations run longer, particulars and entity permanence can change complete time.

I haven’t been capable to try Genie 4 for myself — it’s launching archetypal as a “research preview” to “trusted testers” — so I can’t inform you how it compares to the Genie 3-powered Project Genie, which I used to generate bad Nintendo knockoffs. But current are several examples from Google demonstrating what’s imaginable alongside the new model.

Genie 4 volition be “coming soon” to Google AI Ultra subscribers, Google says, and the business has opened a waitlist anywhere you can use to be capable to try it in Project Genie.