DeepMind unveils Genie 4 world model with native audio and live steering
Google DeepMind announced Genie 4, a world model adding native audio and real-time language steering, now in research preview for trusted testers.
The story in focus
Google DeepMind announced Genie 4, a research-preview world model that adds native audio and lets users steer simulated scenes in real time with natural language. It is rolling out to trusted testers first, with wider access planned for Google AI Ultra subscribers and early Gemini API partners.
- Genie 4 adds on-the-fly audio, including ambient sound, music and interactive dialogue.
- Users can steer scenes mid-simulation with natural language, such as changing weather or the camera.
- Google says scenes can stay consistent for up to two minutes, double Genie 3's limit.
- Access begins with trusted testers; Google AI Ultra and early API partners to follow.
What DeepMind announced
Google DeepMind announced research on Genie 4, a world model for real-time interactive simulation that adds native audio and what the company describes as greater realism. The model builds on capabilities from the Gemini family; DeepMind says Gemini Omni anchors simulations in real-world appearance and behavior, while other Gemini models supply reasoning aimed at keeping a simulation coherent and continuous.
According to the announcement, users can set an event in motion and steer scenes in real time using natural language, such as moving the camera, changing the weather or introducing new characters. Every world also generates audio on the fly, including ambient sounds, music and interactive dialogue.
Access and stated limits
Genie 4 has started rolling out to a limited group of trusted testers. DeepMind says it is coming soon to Google AI Ultra subscribers in Project Genie and to early access partners in the Gemini API and AI Studio. TechCrunch reported that the model is in research preview and that its team acknowledged simulations can drift in detail the longer they run, with audio accuracy and in-world text legibility not yet fully solved.
Why DeepMind says it matters
DeepMind frames world simulation as a step toward AGI, citing safe training environments for robots and agents and creative tools for filmmaking, journalism, gaming and manufacturing. TechCrunch reported a research experiment with a Gemini 4 Argon model in which tasks were broken into steps that Genie 4 simulated in real time, suggesting a possible way to test robot scenarios in simulation.
When will Genie 4 be more widely available?
Google DeepMind said Genie 4 is rolling out first to a limited group of trusted testers. It is coming soon to Google AI Ultra subscribers in Project Genie and to early access partners in the Gemini API and AI Studio.
Sources & further reading
- We’re announcing Genie 4, our new world model for real-time interactive simulation. ↗Google Blog
- Google’s Genie 4 world model adds audio, more realistic worlds, and real-time steering | TechCrunch ↗TechCrunch
- Google’s next Genie AI model generates more realistic experiences ↗The Verge
Help keep the record accurate. Send a correction.
