Genie 4 is now available
Google is rolling out Project Genie, a browser tool for generating and exploring interactive worlds, to eligible Google AI Ultra subscribers. The big additions are Genie 4, which lets you change a world mid-session with new prompts and adds native audio, plus a community gallery where people can publish and remix worlds.
- Genie 4 lets you edit a running world by typing new prompts, such as shifting the weather or adding a character, and it adds real-time audio that adapts to the scene.
- Genie 3 gets sessions extended from 60 seconds to 3 minutes per generation, world chaining, and seeding from 360-degree panoramas.
- A new agent builds worlds from natural-language descriptions and can bring Gemini in to suggest next steps or trigger events, and a gallery lets users remix and publish worlds.
Google has rolled out a new Project Genie experience, the research prototype it first introduced earlier this year for generating and exploring interactive worlds in the browser. The update is aimed at eligible Google AI Ultra subscribers, with Google AI Pro members able to join a waitlist. The headline addition is Genie 4, which Google calls its most realistic and interactive world model yet. The big change is that you can type new prompts mid-session to alter a running world: shift the weather, drop in a character, redirect the action. Genie 4 also adds real-time native audio that adapts as the scene plays out, from ambient soundscapes to character dialogue, which is a first for the line.
Genie 3 hasn't been retired. It stays in the model picker for situations where you want something Genie 4 doesn't offer yet, and Google has upgraded it in three ways. Sessions now run up to 3 minutes per generation, up from 60 seconds, and you can chain worlds together for longer journeys. You can seed new worlds from 360-degree panorama images. And you can trigger in-world events with natural-language prompts while exploring. Google also introduced a Project Genie agent that takes a plain-language description and orchestrates its multimodal models to produce the initial visuals and a custom background score. Inside a session, Gemini can watch what's on screen, offer context-aware suggestion chips for where to go next, or trigger events on its own.
On the social side, there's a new community gallery mixing curated worlds with user creations. Examples include "Living Memory," which turns personal photos into explorable spaces, "Dino Field Guide," built on 360-degree prehistoric panoramas, and "Rock Paper Scissors," where Gemini referees a desk showdown between objects. Google also highlights two collaborations with outside creators: "Villa Giallo," a room-by-room gothic mystery made with creative technologist Will aka BLVCKL!GHT, and "The Life of Ants," a filmmaker Anand Gandhi project that follows a colony from an ant's perspective using Genie 4's photorealism. Users can remix gallery worlds by changing structure, soundtrack, or rules, then publish their own. Google says all generated video and audio carries SynthID watermarking and that inputs and outputs pass safety screening. The post admits open problems in camera latency, complex physical interactions, real-time audio precision, and keeping scenes consistent over long stretches, and points to a separate Genie 4 research overview for details.