Google DeepMind releases Genie 4 world model in research preview with audio and real-time scene steering
Google DeepMind released Genie 4, its newest world model, in research preview on 27 October 2026. It generates more photorealistic, interactive worlds than its predecessor and adds real-time audio (music, ambient sound, dialogue). Users can type prompts mid-simulation to change weather, camera angle, people or objects. Scenes stay consistent for up to two minutes, double the Genie 3 limit. Gemini now acts as a planner and frame checker alongside Genie. Google lists gaming, interactive learning, robotics simulation and film prototyping as use cases. The article describes no security incident, vulnerability or threat activity.
Detalhes técnicos
- Access is limited to trusted testers. Rollout to Google AI Ultra subscribers in Project Genie, and to early access partners of the Gemini API and Google AI Studio, is planned in the coming weeks.
- A Gemini Robotics experiment used a Gemini 4 Argon model to break tasks into steps that Genie 4 simulated in real time.
- Known limitations: details and objects drift over longer runs, and audio accuracy and in-world text legibility are unresolved.
- Market context: Runway released GWM Worlds 2, AMD is acquiring World Labs, Amazon is reportedly in talks to buy Decart, and Rivian acquired Third Dimension.
Linha do tempo
-
Google DeepMind releases Genie 4 world model in research preview with audio and real-time scene steering
ler a análise desta fonte ocultar
Google DeepMind released Genie 4, its newest world model, in research preview on 27 October 2026. It generates more photorealistic, interactive worlds than its predecessor and adds real-time audio (music, ambient sound, dialogue). Users can type prompts mid-simulation to change weather, camera angle, people or objects. Scenes stay consistent for up to two minutes, double the Genie 3 limit. Gemini now acts as a planner and frame checker alongside Genie. Google lists gaming, interactive learning, robotics simulation and film prototyping as use cases. The article describes no security incident, vulnerability or threat activity.
Technical Details
- Access is limited to trusted testers. Rollout to Google AI Ultra subscribers in Project Genie, and to early access partners of the Gemini API and Google AI Studio, is planned in the coming weeks.
- A Gemini Robotics experiment used a Gemini 4 Argon model to break tasks into steps that Genie 4 simulated in real time.
- Known limitations: details and objects drift over longer runs, and audio accuracy and in-world text legibility are unresolved.
- Market context: Runway released GWM Worlds 2, AMD is acquiring World Labs, Amazon is reportedly in talks to buy Decart, and Rivian acquired Third Dimension.