Genie 4: Google DeepMind's world model you steer live, with audio
Why it matters
Google DeepMind announced Genie 4 on 27 October 2026, its new world model for real-time interactive simulation. Describe a scene, then explore it and steer it live in natural language: move the camera, change the weather or introduce new characters. Every world also generates native audio on the fly, including ambient sounds, music and interactive dialogue. Google says it builds on capabilities from the Gemini family, with Gemini Omni anchoring realistic appearance and behavior, and pitches it for safe training environments for robots and agents and for creative tools. Google lists limits: visual details and object permanence can drift over time, and it supports a few minutes of continuous interaction. It is rolling out to a limited group of trusted testers, coming soon to Google AI Ultra subscribers in Project Genie and early access partners in the Gemini API and AI Studio; all Genie 4 video and audio carries SynthID.
Read the source: blog.google (Google DeepMind) (opens in a new tab)
All hits from 27 Oct 2026More ModelsGet the daily hits by email