Science & Tech · AI · 4 hrs ago
Google DeepMind previews Genie 4, an interactive AI world model
Google DeepMind has introduced Genie 4, an AI model that creates virtual worlds people can change while exploring them.
Unlike a finished video, the simulation responds to users in real time: they can change the weather, move the camera or add people and objects with text prompts.
It can also generate matching sounds, music and character dialogue as events unfold.
Google says the model is more realistic than its predecessor, Genie 3, and can keep a scene consistent for up to two minutes, though another account describes use as lasting a few minutes.
The company sees possible uses in games, education, film-making and robotics, including testing robot tasks in simulation.
But objects can change appearance or position during longer sessions, and accurate audio and complex interactions remain challenges.
Genie 4 is currently in a limited research preview, with Google planning access for Google AI Ultra subscribers and developers; generated video and audio will carry SynthID watermarks.
Google DeepMind introduced Genie 4, a world model that generates interactive virtual environments in real time.
The model can add audio, including environmental sounds, music and dialogue, and lets users change scenes with text prompts.
Google released Genie 4 as a research preview for a limited group of testers.
The model works with Gemini to plan and monitor actions, and Google is testing its use in robot simulations.
Google says Genie 4 could support gaming, filmmaking, education and robotics, but long simulations can lose consistency and audio accuracy remains a challenge.
- Who
- Google DeepMind introduced Genie 4.
- What
- It is an interactive world model that generates virtual environments and audio in real time.
- When
- October 27, 2026.
- Where
- Not stated.
- Why
- Google says the model could be used for gaming, filmmaking, education and robotics.
This story does not have two clearly opposing sides.
Google released Genie 3 as a research preview.
Runway released its second world model, GWM Worlds 2.
Google DeepMind introduced Genie 4 as a research preview.
- Model
- Genie 4
- Release status
- Research preview for a limited group of testers
- Audio
- Generates music, ambient sounds and interactive dialogue
- Scene consistency
- Up to two minutes, according to TechCrunch
- Planned access
- Google AI Ultra subscribers and developers are expected to get access later
4 sources
What people are saying
What we’ll increasingly see in the future are these hybrid world models, where you have a combination of Gemini writing code and text, and Genie ingesting that in real time and simulating reality at different levels of abstraction
The real time aspect becomes most valuable in the areas where keeping a state of flow in the creativity space becomes incredibly useful