
Google Makes Genie 4 Available to AI Ultra Subscribers Through Project Genie
Google has started rolling out Genie 4 through Project Genie, giving eligible Google AI Ultra subscribers access to its latest interactive world model. The November 18 update also introduces a conversational creation agent, a community gallery for sharing and remixing virtual worlds, and new capabilities for the existing Genie 3 model.
Genie 4 is now available in Project Genie.
It brings native audio, real-time editing, and a new agent to help you create interactive, virtual worlds. pic.twitter.com/CipplkOuMP
– Google DeepMind (@GoogleDeepMind) November 18, 2026
The release marks the first availability of Genie 4 inside Project Genie, Google’s experimental platform for generating and exploring interactive environments. Users can now change elements of a world while exploring it, add characters through text prompts, and hear audio generated in real time as scenes unfold.
Google AI Pro subscribers can join a waitlist for access, while the initial rollout remains limited to eligible Ultra members.
Genie 4 Brings Real-Time Editing and Native Audio
The main addition is Genie 4, which gives users more control over the environments they generate. Rather than relying only on an initial text prompt or image, users can enter new instructions during a session to change the world around them.
For example, a prompt can shift the weather, introduce a new character, or redirect the action. The model then adjusts the environment as the user continues exploring, creating a more interactive experience than a fixed sequence of generated scenes.
Genie 4 also introduces native audio generation. Its environments can include ambient sounds that match the setting, along with music and character dialogue that responds to events.
Google describes Genie 4 as its most realistic and interactive world model so far. The model combines capabilities from the Gemini family to generate environments, interpret instructions, and maintain interactions as scenes change.
Project Genie Agent Turns Conversations Into Interactive Worlds
Google has also introduced a new Project Genie agent that makes world creation less dependent on writing a detailed prompt from scratch.
Users can describe an idea in natural language, and the agent coordinates Google’s multimodal models to build the initial environment. This process includes generating the visuals that establish the world’s appearance and creating background music suited to its atmosphere.
Gemini models can also act as companions during exploration. They observe the current experience, suggest relevant actions, and can trigger events within the world. This lets users shape the environment on the fly without pausing the session.
New Community Gallery Supports Sharing and Remixing
The update adds a community gallery where users can browse curated experiences, explore worlds created by others, and remix existing projects.
Google has highlighted several examples to demonstrate the range of experiences available. Villa Giallo, created with creative technologist Will, known as BLVCKL!GHT, turns a gothic manor into an interactive mystery that visitors can explore room by room.
Another featured experience, The Life of Ants, was developed with filmmaker Anand Gandhi. It places users inside a detailed ant colony, allowing them to explore the environment from an ant’s perspective.
Users can also publish their own creations to the gallery, giving others the opportunity to explore and modify them.
Genie 3 Gets Longer Sessions and 360-Degree Panorama Support
Genie 3 remains available through Project Genie’s model selector, with several upgrades aimed at longer and more continuous exploration.
The updated model supports sessions of up to three minutes per generation, compared with the previous 60-second limit. Users can also connect multiple worlds to create longer journeys instead of ending an experience after a single generated environment.
Another addition is 360-degree panorama image grounding. Users can start an environment from a panoramic image, giving the model a visual reference for the surrounding space.
Genie 3 also gains real-time in-world events controlled through natural-language prompts. This allows users to change elements of an environment while exploring.
The two models serve different purposes within Project Genie: Genie 4 introduces native audio and more advanced interactive world generation, while Genie 3 remains an option for longer sessions and panorama-based experiences.
Who Can Access Genie 4?
The new experience is rolling out to eligible Google AI Ultra subscribers in regions where the subscription is available. Users must be at least 18 years old, with access rolling out gradually by region.
Google AI Pro subscribers can join the waitlist through the Project Genie website. Google has not announced a specific date for broader Pro access.
Daily usage limits vary by subscription tier and server load, and users can expect brief wait times during peak demand.
Those without creation access can still view short previews of experiences in the gallery. Google says generating worlds during the current early-access period does not use AI credits.
Is Genie 4 Available Through a Public API?
Google’s November 18 update focuses on Project Genie subscribers and does not include a general API release for developers.
An earlier announcement on October 27 identified selected early-access partners through the Gemini API and Google AI Studio. That developer access is separate from the new subscriber rollout, and Google has not announced general availability for all developers.




