Google's newest world model builds worlds you can hear and talk to.
Google DeepMind released Genie 4 on Tuesday, an AI that builds explorable virtual worlds in real time. It is DeepMind's first world model that creates its own music, background sounds and spoken dialogue. Trusted testers can use it now, and Google AI Ultra subscribers get it in the coming weeks.
Why it mattersYou can type a request mid-scene, such as rain or a new character, and the world changes. Scenes now stay consistent for up to two minutes, twice as long as with Genie 3.
▶ Related video: Genie 4: A new generation of world simulation
Read in the app — free, in 9 languages
Related stories
OpenAI's new mid-size model costs half as much as the one before.2026-10-27Block's new voice AI learned to talk without reading any text.
2026-10-25Moonshot's new Kimi trails only three models on an independent index.
2026-10-24A free Chinese image model passed a million downloads in a week.
2026-10-23Canva trained its first image model on templates, not photos.
2026-10-23