Google just cracked open access to one of its most ambitious AI experiments yet. Starting today, Google DeepMind is rolling out Project Genie to Google AI Ultra subscribers in the U.S., giving paying customers the ability to create, explore, and remix interactive worlds using nothing but text prompts and images. Powered by the Genie 3 world model previewed last August, the prototype represents a major leap in generative AI's evolution from static content creation to dynamic, navigable environments that generate in real time as you move through them.
Google is making a bold bet that the future of AI isn't just about generating images or text, but entire interactive universes you can step inside. Project Genie, now available to Google AI Ultra subscribers in the U.S., is the company's experimental research prototype that transforms world-building from a developer's domain into something anyone can do with a few words and a sketch.
The launch represents the next evolution of Genie 3, Google DeepMind's general-purpose world model that's been impressing trusted testers since its August preview. Unlike static 3D environments or pre-rendered game worlds, Genie 3 generates the path ahead in real time as you move and interact. "Even in this early form, trusted testers were able to create an impressive range of fascinating worlds and experiences, and uncovered entirely new ways to use it," Product Manager Diego Rivas wrote in today's announcement.
The technology addresses a fundamental challenge in building artificial general intelligence: creating systems that can navigate the messy, unpredictable diversity of the real world. While Google DeepMind has built specialized agents for constrained environments like , Genie 3 simulates physics and interactions for dynamic worlds with what the company calls "breakthrough consistency." That consistency enables simulation of any real-world scenario, from robotics testing to historical recreations to pure fiction.
