Google DeepMind releases Genie 4 in Project Genie for AI Ultra subscribers, adds native audio
Genie 3 gets sessions of up to 3 minutes, up from 60 seconds, and Google AI Pro members can join a waitlist for the new experience.
What happened
Google DeepMind has released Genie 4, an AI model that generates interactive worlds you can explore in your browser. It is called a world model. You can type new prompts while inside a world to change the weather or add a character. Genie 4 also generates sound that adapts to the scene.
Why it matters to you
It shows AI moving from making still images and video toward making spaces people can step into and change. For now it is a research prototype open to paying subscribers.
The takeaway
If you subscribe to Google AI Ultra, you can try building and editing AI-generated worlds today.
What it means for you
Genie 3 stays available for longer runs and panorama-seeded worlds, while Genie 4 offers the highest fidelity visuals, physics and native audio.
Creators can remix gallery worlds and publish their own, and Google showcases collaborations with creators such as filmmaker Anand Gandhi.
Access is limited to Google AI Ultra subscribers, with Google AI Pro members on a waitlist, which ties the world model work to Google's paid subscription tiers.