Google is updating Project Genie, an experimental research prototype for creating and exploring interactive worlds in a browser. Eligible Google AI Ultra subscribers can now access the Genie 4 world model for the first time, alongside upgraded Genie 3 capabilities, a creation assistant and a community gallery.
According to Google, the changes partly reflect feedback from early testers, who wanted more control over events within worlds and easier ways to create and share them. Project Genie combines Google’s world models with its Gemini models.
Genie 4 and the upgraded Genie 3
With Genie 4, users can alter an environment while exploring it by entering new text prompts. They can, for example, change the weather, add a character or redirect the action. Alongside the visuals, the model provides natively generated real-time audio that adapts to the scene, including ambient sounds and character dialogue.
Google says Genie 4 offers its highest-fidelity visuals and physics, as well as native audio, but Genie 3 remains available for users who want longer runs or spaces based on panoramic images. The Genie 3 update allows exploration for up to three minutes per generation, up from 60 seconds, and lets users link multiple worlds into longer journeys. It can also create worlds based on 360-degree panorama images, while events can be changed during exploration using natural-language prompts.
Conversational creation and a community gallery
The new Project Genie agent coordinates Google’s multimodal models to help users realize an idea described in natural language. Alongside the initial visual style of a world, it creates a custom background score suited to the mood. Gemini models can also be brought into an exploration session: they observe what is happening on screen, suggest context-aware next steps or automatically trigger events in the world.
The new gallery contains both curated and community-created experiences. “Living Memory” turns personal photos into explorable spaces, “Dino Field Guide” offers prehistoric habitats based on 360-degree panoramas, and “Rock Paper Scissors” pits themed objects against one another on a desk, with Gemini acting as a real-time referee.
Featured collaborations include “Villa Giallo,” made with creative technologist Will aka BLVCKL!GHT, a room-by-room mystery set in a gothic manor. “The Life of Ants,” created with filmmaker Anand Gandhi, presents an ant colony from an ant’s perspective. Both use Genie 4.
Existing gallery worlds can be remixed with the agent’s help, allowing users to change their structure, soundtrack or rules. Users can also publish their own creations for others to explore and remix.
Safety and availability
Google says all video and audio generated in Project Genie carries SynthID watermarking, and model inputs and outputs undergo safety screening. The company identifies camera latency, complex physical interactions, real-time audio precision and long-term scene permanence among the remaining research challenges.
The new Project Genie experience is available from the announcement to eligible Google AI Ultra subscribers. Google AI Pro members can join a waitlist.



