Google opens its Genie 4 world model to AI Ultra subscribers
Google is rolling out Genie 4 to eligible Google AI Ultra subscribers in Project Genie, its browser-based world-building prototype, along with a creation agent and a community gallery. Genie 4 lets users change worlds mid-session and adds native audio, though Google says details still drift in long sessions. Google has opened Genie 4, its newest […] This story continues at The Next Web
Genie 4 creates a world you can move through, and lets you change it as you go. Typing a prompt mid-session can shift the weather, add a character or redirect what happens. It also adds sound for the first time, from background noise to character dialogue.
Google DeepMind first showed Genie 4 to testers in October . The new release adds an AI agent that builds worlds from a conversation, and a gallery where users can share and remix each other’s worlds. Project Genie already lets some users explore simulated cities built from Street View.
The older Genie 3 stays available for longer runs, and now lasts up to three minutes per generation, up from 60 seconds. Big game studios remain wary, and Take-Two says GTA 6 used no generative AI.
Google says both models are still research systems. Details can drift in long sessions, and readable text and complex physics remain hard. Everything made in Project Genie carries SynthID, Google’s invisible watermark for AI content.
World models are drawing big money: AMD agreed to buy Fei-Fei Li’s World Labs for $8.2B in September. Google DeepMind says such models could also train robots and AI agents before they work in the real world. Rival lab Odyssey says its Odyssey-3 model can already control robots and cars.