In brief
- Google is rolling out Genie 4 to eligible AI Ultra subscribers through Project Genie, after announcing the model in October.
- Users can change generated worlds during exploration, with AI-created ambient audio and interactive character dialogue.
- Project Genie also adds an AI creation assistant, a community gallery, and longer sessions for the older Genie 3 model.
Google is opening Genie 4 to eligible AI Ultra subscribers, letting people create digital worlds they can walk through, change with a text prompt, and even populate with characters that talk back.
The company announced Wednesday that its latest world model is rolling out through Project Genie, a browser-based experiment for generating interactive environments. Genie 4 was unveiled in October, but access until now had been limited to trusted testers.
Instead of rendering a finished video from a prompt, Genie generates the scene as someone explores it. With the fourth model, users can type instructions midway through a session to change the weather, introduce a character, or steer what's happening without starting over.
#1 Searched Person on Google 2026?
Genie 4 also creates sound alongside the visuals, from wind and footsteps to dialogue that reacts to events in the scene. Google says the model improves the way people move and objects respond to physical forces, though those claims come from its own demonstrations rather than independent testing.
Now, the not so nice thing you need to know. The update is part of the Google AI Ultra subscription, which costs $199.99 per month in the United States. Google says eligible Ultra users can access the new experience now, while AI Pro subscribers can join a waitlist.
A genie that keeps getting better
Genie has come a long way from the side-scrolling experiments Google showed in 2024. The first version learned from more than 200,000 hours of gameplay footage and could turn a single image into a simple playable 2D environment, as Decrypt reported at the time.
Genie 2 moved the research into three-dimensional spaces later that year, allowing people or AI agents to navigate generated environments with keyboard and mouse controls. Genie 3 took another step toward continuous, real-time exploration, and Project Genie opened in January to adult Ultra subscribers in the United States.
Those early sessions lasted just 60 seconds, and character controls or physical behavior could be unreliable. Google is now keeping Genie 3 available alongside the new model, with sessions extended to three minutes, support for 360-degree panorama images, and the ability to connect multiple generated worlds into longer journeys.
That three-minute limit applies to the updated Genie 3 experience; Google did not specify a comparable session length for Genie 4 in Wednesday's announcement. Users can choose between the models depending on whether they want Genie 4's stronger visuals and audio or Genie 3's longer explorations.
The technology is not the same as a traditional game engine that builds a stable level with fully defined rules. Genie predicts what a world should look and sound like as it changes.
Google acknowledges that objects can drift or change between views, complex physics remain difficult, and generated speech or text may be inaccurate. It says Project Genie's audio and video carry SynthID watermarks and that inputs and outputs undergo safety screening.
Google has not announced when Genie 4 will become available to AI Pro subscribers beyond the new waitlist.
Daily Debrief Newsletter
Start every day with the top news stories right now, plus original features, a podcast, videos and more.

