Skip to content
← All briefings

Briefing · Oct 27

Google DeepMind unveils Genie 4, a photorealistic interactive world model built on Gemini Omni

AI

Google DeepMind unveils Genie 4, a photorealistic interactive world model built on Gemini Omni

Google DeepMind launched Genie 4, its most advanced interactive world model to date, describing it as capable of generating dynamic environments that respond in real time to user input — with realistic physics, sound, and characters that move, react, and speak. The model is built on Gemini Omni for lifelike visual detail and supports on-the-fly scene editing, making it a significant step toward AI-generated interactive worlds.

DeepMind outlined three headline advances: photorealism across a wide range of environments and subjects, advanced physics simulation including aerodynamics and natural human movement, and real-time prompt-driven controls that let users add elements or alter the scene mid-session.

This briefing is part of the archive.

Get the app for full archive access — every daily briefing, sourced and attributed.