Genie 4 adds sound to worlds you can reshape in real time

Google DeepMind’s new Genie 4 world model pairs environments you can explore and reshape with sound generated in real time. Announced on October 27, 2026, it is currently rolling out as a research preview for a limited group of trusted testers.

Change the weather while the scene keeps moving

In DeepMind’s demonstrations, users navigate and control the camera with standard keyboard controls. Natural-language prompts let them change the weather, move the camera or introduce a person or object while the simulation runs.

Genie 4 generates ambient sounds, music and interactive dialogue alongside the visuals. DeepMind says characters can converse and react to user input or events within the environment, so a scene’s sound can evolve with what happens in it.

Gemini Omni helps ground the world’s appearance and behavior. Other Gemini models help maintain continuity as events unfold and users intervene.

Trying an idea before writing engine code

Trusted testers in game development and creative technology are experimenting with explorable prototypes, using prompts to try visual styles, materials and interactive mechanics before writing engine code. These are early concept tests, not demonstrations of complete commercial game production.

Filmmakers are also testing Genie 4 for previsualization: stepping into a simulated scene and adjusting camera angles, lighting, dialogue and action across different takes.

An object may not stay the same

DeepMind says Genie 4 can keep environments coherent over short sessions, but visual details and object permanence can drift. An object may change or fail to remain consistent as a session gets longer, users move through the environment or repeated complex prompts alter the scene.

Complex physical interactions, accurate real-time audio, readable text, spatial accuracy and long-term consistency remain research problems. DeepMind also says generated video and audio will include SynthID watermarks to help identify AI-generated content.

Who gets access next

DeepMind says Genie 4 is coming soon to Google AI Ultra subscribers through Project Genie, and to early-access partners through the Gemini API and Google AI Studio. That does not mean every Ultra subscriber can use it today.

A Project Genie waitlist is available, but joining it does not provide immediate access. No dated general release has been announced.