Genie 4 is now available
Reported by Foro3D, Blockchain News, Google DeepMind and 4 more outlets7 sources ↓
Google is rolling out Genie 4 in Project Genie to eligible Google AI Ultra subscribers. The update adds real-time scene edits, native audio, a creation agent and a gallery for sharing and remixing worlds; Google AI Pro members can join a waitlist.
Key points
- Users can change a world with prompts while exploring, including its weather, characters or events
- Genie 4 generates native audio, including ambient sound and character dialogue
- A conversational agent helps create worlds, and a gallery lets users publish and remix them
- Genie 3 remains available for longer runs and panorama-seeded spaces; its generation limit is now three minutes
- Google has not stated Genie 4 session length or regional availability
Sources
- Foro3DRead ↗DeepMind's Genie 4: native audio and editable scenes in real time
- Blockchain NewsRead ↗Genie 4 launches in Project Genie with audio and live edits
- Google DeepMindRead ↗Genie 4 is now available
- TipRanksRead ↗Google says Genie 4 now available
- Unite.AIRead ↗Google Brings Genie 4 to Project Genie for Eligible AI Ultra Users
Details
The rollout follows Genie 4’s October 27 research-preview announcement for trusted testers. Google’s November 18 announcement makes the model available through Project Genie to eligible Google AI Ultra subscribers; Google AI Pro members can join a waitlist, but are not in the initial rollout.
Google says Genie 4 can respond to prompts during a session, letting users change elements such as weather, characters or the direction of events without starting over. It also generates real-time audio that adapts to the scene, including ambient sound and character dialogue.
The update adds an agent that creates worlds from natural-language descriptions and a gallery where users can publish and remix worlds. Genie 3 remains selectable for longer runs or spaces seeded from 360-degree panoramas; Google says its generation limit has increased from 60 seconds to three minutes.
Google says generated video and audio include SynthID watermarks and that model inputs and outputs undergo safety screening. It identifies camera latency, physical interactions, real-time audio precision and long-term scene permanence as ongoing challenges.
Written from coverage as of
Images
- Google DeepMind
- The Decoder
- Unite.AI
- Tech Times
Stocks in this story
Change since open: 20 Nov to 21 Nov 2026
Last refresh