Google Brings Genie 4 to AI Ultra Users: Access, Features & Limits
Google has opened Genie 4 to eligible AI Ultra subscribers in Project Genie for the first time, adding live audio, mid-session edits and a creation agent.
Image: Wikimedia Commons / National Academies - Earth and Life Studies
Google opened first access to Genie 4, its most realistic and interactive world model, to eligible Google AI Ultra subscribers on November 18, 2026, moving the system out of a research preview for trusted testers and into a consumer product. The rollout, announced by Google DeepMind, adds a creation agent and a community gallery to Project Genie, the browser prototype where users generate and walk through interactive worlds.
Genie 4 changes a world while the user is inside it. A mid-session prompt can shift the weather, add a character or redirect the action, and the model now generates native audio — ambient sound, music and dialogue — that reacts as the scene unfolds. Google DeepMind calls Genie 4 natively multimodal.
Key Takeaways
- Google opened Genie 4 to eligible AI Ultra subscribers on November 18, 2026 — first consumer access after an October 27 preview for trusted testers.
- Users can prompt mid-session to change weather, characters and events, and Genie 4 generates native audio — ambient sound, music and dialogue — for the first time.
- A new Project Genie agent builds worlds from natural-language conversation, and a community gallery lets users explore, remix and publish shared worlds.
- Google AI Pro members can join a waitlist; Genie 3 stays available, with sessions extended to three minutes, up from 60 seconds.
What Is Genie 4 and How Does It Work in Project Genie?
Genie 4 is a world model: it draws the next moment of a scene frame by frame as the user moves and types, with no game engine and no stored 3D map underneath. Gemini models work inside the session: Gemini Omni supplies a real-world reference for how simulations look and behave, and Gemini 3.8 Flash acts as a guiding agent, watching the scene and splitting goals into steps.
Every clip generated in Project Genie carries SynthID, Google’s imperceptible watermark for AI-generated content, and all inputs and outputs pass through safety screenings. Google DeepMind is open about the limits: camera latency, intricate physical interactions, real-time audio precision and long-term scene permanence remain unresolved, and details can drift in longer sessions.
How to Get Access to Genie 4 in Project Genie
Eligible Google AI Ultra subscribers can try the Project Genie experience from November 18, 2026, and Google AI Pro members can join a waitlist on the same page. Google has announced no general availability date for Genie 4 beyond that rollout, and no API price or model identifier. The announcement is on Google’s official blog.
Project Genie began on January 29, 2026, with Genie 3 for AI Ultra subscribers in the United States. Google DeepMind announced Genie 4 on October 27 as a research preview for trusted testers; the November 18 rollout carries out the Project Genie part of that plan, following the rollout of Gemini 4 and DeepMind’s AlphaGenome 2.
What Is New in Genie 3 and the Project Genie Gallery?
Genie 3 stays in Project Genie’s model picker for longer runs or worlds seeded from 360-degree panorama images. Sessions now run up to three minutes per generation, up from 60 seconds, and users can link worlds into longer journeys or trigger real-time events with natural-language prompts.
The gallery gathers curated worlds and community work: Living Memory turns photos into explorable spaces, Dino Field Guide explores prehistoric habitats, and Rock Paper Scissors uses Gemini as a real-time referee. Creator collaborations include Villa Giallo and The Life of Ants. Users can remix gallery worlds and publish their own versions.
| Feature | Genie 3 | Genie 4 |
|---|---|---|
| First access | January 29, 2026 | November 18, 2026 |
| Native audio | No | Yes: sound, music, dialogue |
| Mid-session edits | Added in update | Built in |
| Session length | Up to 3 minutes | A few minutes |
Frequently Asked Questions
How can I try Google Genie 4?
Eligible Google AI Ultra subscribers can open the new Project Genie experience from November 18, 2026. Google AI Pro members can join the waitlist on the same page.
How long can a Genie 4 world last?
Google DeepMind describes a few minutes of continuous interaction. Details and object permanence can drift in longer sessions, and scene permanence remains an active research challenge.
What does the Project Genie agent do?
It builds a world from a natural-language conversation, creating opening visuals and a custom background score. Gemini can also suggest next steps and trigger in-world events while the user explores.
Google has set no date for general availability of Genie 4 in the Gemini API; both Genie models remain frontier research systems.