Google on November 18, 2026 opened a redesigned Project Genie to eligible Google AI Ultra subscribers, and with it the first public access to Genie 4, its newest interactive world model. Project Genie is an experimental research prototype that lets people generate worlds and move through them from a browser. In an announcement signed by Diego Rivas of Google DeepMind and Pavan Kumar of Google Labs, the company says early testers asked for more control inside the worlds and easier ways to build and share them.

Genie 4 is described as Google's most realistic and interactive world model so far. Its main change for users is mid-session prompting: while navigating a world, a person can type a new prompt to alter the environment in real time, for example by changing the weather, adding a character or redirecting the action. Google also says Genie 4 pairs its visuals with real-time native audio for the first time, from ambient sound to character dialogue, and that the audio adapts as the scene develops.

Genie 3 stays available in the model picker, and Google is upgrading it for people who want longer runs or panorama-seeded spaces. Sessions can now last up to three minutes per generation, up from 60 seconds, and several worlds can be chained into a longer journey. Genie 3 can also start from 360-degree panorama images, and it supports real-time in-world events triggered by natural-language prompts.

The release adds two new layers around the models. The Project Genie agent lets a user describe an idea in conversation. It coordinates Google's multimodal models to produce an opening look for the world and a custom background score, and during a session it can bring in Gemini models that watch the screen, suggest where to go next or trigger in-world events. A community gallery collects curated worlds and creations from users, who can publish their own and remix others by changing the structure, soundtrack or rules. Google highlights examples such as "Living Memory," "Dino Field Guide," "Rock Paper Scissors," a gothic-manor mystery called "Villa Giallo" made with creative technologist Will aka BLVCKL!GHT, and "The Life of Ants," built with filmmaker Anand Gandhi.

Availability is limited. Eligible Google AI Ultra subscribers can use the new experience starting November 18, and Google AI Pro members can join a waitlist on the same page. Google does not give countries or pricing in the post.

Google also lists safeguards and limits. Video and audio generated in Project Genie carry SynthID watermarking, and all model inputs and outputs go through safety screening. The company calls Genie 3 and Genie 4 world models in the frontier research space and names open challenges: camera latency, intricate physical interactions, real-time audio precision and long-term scene permanence. It points readers to a separate Genie 4 model research overview for details on limitations.

The announcement contains no benchmark figures or independent evaluation. The claims about realism and interactivity are Google's, and the post does not state how long a Genie 4 session can run.

References