Google's AI worlds just got a lot more alive. On 18 November 2026, Google began rolling out a new Project Genie experience to eligible Google AI Ultra subscribers, with first-time access to Genie 4, its most realistic and interactive world model yet. Genie 4 was unveiled on 27 October as a research preview for trusted testers — this is the first time paying subscribers can walk through its worlds.
Project Genie is Google's experimental tool for creating and exploring interactive virtual worlds from text prompts. Five changes in this update matter most.
1. Change the world while you walk through it
The biggest change is mid-session prompting: with Genie 4, you can type a new prompt while moving through a world you created — change the weather, add a character, shift the camera — without restarting. Google's demo starts with an astronaut on a spacewalk above Earth, then adds a giant alien spaceship passing by, mid-scene.
2. The world now has sound — and voices
Genie 4 is the first Genie model to pair its visuals with real-time native audio that adapts as the scene unfolds: ambient soundscapes, music and character dialogue generated alongside the pictures, reacting to what happens — wind, footsteps, a voice answering back.
Sound does most of the work of making a place feel real: a silent world is a picture you can walk through, while a world with sound that responds to you is a place. Google has not said how precise the audio is, and lists real-time audio precision as an open challenge — so expect rough edges, especially in dialogue.
3. An agent helps you build the world
The new experience adds a Project Genie agent, which helps users build their vision through conversation. Instead of writing one perfect prompt, describe what you want in chat — Google's example builds a point-of-view ski run down a snowy alpine course, then adds an avalanche — and the agent turns it into a world.
World-building by prompt is hard: one prompt must carry the setting, style, camera and mood. An agent that asks questions and refines the idea lowers the skill needed.
4. A community gallery to explore and remix
The update adds a community gallery to discover, explore and remix other people's worlds. Google's demo picks a downhill ski adventure from the gallery and remixes it with one prompt — "show me this downhill course at night" — turning day into night.
The gallery solves the cold-start problem: a new user faces a shelf of worlds to change, not a blank prompt box.
5. Genie 3 gets longer sessions and new tricks
The update is not only about Genie 4: Genie 3 stays in the model picker for longer runs and gets its own upgrades — sessions of up to three minutes, up from 60 seconds, plus world chaining, 360-degree panorama grounding and real-time in-world events. Google's example is a wizard riding an orange dragon through a misty forest.
Three minutes is still short, but it triples exploration time before a reset. Genie 3 remains the model for longer sessions, Genie 4 the one for editing and sound; Google has not stated Genie 4's session length.
Who can use it, and what it costs
Access starts with eligible Google AI Ultra subscribers, who can try the new Project Genie experience from today; Google AI Pro members can join a waitlist. Ultra costs $199.99 a month in the US. Google has not announced when Genie 4 will reach Pro subscribers beyond the waitlist.
Two limits are worth knowing. First, both Genie 3 and Genie 4 remain frontier research world models: Google lists camera latency, intricate physical interactions, real-time audio precision and long-term scene permanence as open challenges, so worlds can drift and objects may not stay where you left them. Second, all video and audio generated in Project Genie carries a SynthID watermark, and inputs and outputs go through safety screening.
Key takeaways
- Google rolled out Genie 4 in Project Genie on 18 November 2026 to eligible AI Ultra subscribers; AI Pro members can join a waitlist.
- Genie 4 allows mid-session prompting: type a new prompt while exploring to change the weather, add characters or redirect the action.
- It is the first Genie model with real-time native audio — ambient sound, music and character dialogue.
- A Project Genie agent helps build worlds from a description, and a community gallery lets users discover and remix other people's worlds.
- Genie 3 was upgraded too: sessions up to three minutes, world chaining, 360-degree panorama grounding and real-time in-world events.
- Both models remain frontier research, with open challenges in camera latency, physical interactions, audio precision and scene permanence; all output carries a SynthID watermark.
Conclusion
Genie 4 is not a new kind of model so much as a more usable one: the same world-building idea, now editable as you explore it, with sound, a helper agent and a gallery to start from. The price — $199.99 a month — and the research-prototype limits mean it is still an early-adopter toy. But the direction is clear: AI worlds are moving from clips you watch to places you can change, and Genie 4 is Google's clearest step in that direction yet.
FAQ
What is Google Genie 4?
Genie 4 is Google DeepMind's most realistic and interactive world model, creating explorable virtual worlds from text prompts. It rolled out in Project Genie on 18 November 2026, adding mid-session editing and real-time native audio.
How can I try Genie 4?
Genie 4 is available to eligible Google AI Ultra subscribers in Project Genie; AI Pro members can join a waitlist. Ultra costs $199.99 a month in the US, and Google has not announced when Genie 4 will reach Pro subscribers.
What is the difference between Genie 3 and Genie 4?
Genie 4 adds mid-session prompting and real-time native audio, while Genie 3 remains the model for longer sessions — up to three minutes, with world chaining, 360-degree panorama grounding and real-time in-world events.
Sources
- Google Blog — Project Genie update featuring Genie 4
- Decrypt — Google's Genie 4 Is Here, Letting Ultra Subscribers Build AI Worlds in Real Time
- MLQ.ai — Google DeepMind releases Genie 4, a world model users can change from inside the scene
- Daily Tech & AI — Google Sovereign AI Accelerator: How to Apply
- Daily Tech & AI — Devin Bots vs Dots vs Duet: AI Coding Managers
- Daily Tech & AI — AI Tools Collection
Researched with AI assistance and reviewed for accuracy.
Follow on WhatsApp
Follow Daily Tech & AI on WhatsApp for one tech and AI update every day: Daily Tech & AI WhatsApp channel.
Comments
Post a Comment