Google’s Genie 4 AI model creates more realistic worlds with audio and real-time updates

Google has released Genie 4, its latest world model AI, as a research preview to select testers, according to TechCrunch and The Verge.

Genie 4 generates highly photorealistic and interactive virtual worlds that include real-time audio such as music, ambient sounds, and dialogue, enhancing the immersive experience.

The model supports free-form edits during simulations, enabling users to change weather, camera angles, or add new characters and objects on the fly.

It can maintain a consistent scene for up to two minutes, doubling the previous model's limit, and features improved physics and more humanlike characters.

Google integrates its Gemini AI model with Genie 4 to direct scene planning and frame checking, which has been demonstrated in robotics simulations to test tasks before real-world execution.

Google envisions Genie 4 being useful for filmmakers' pre-visualization, game prototyping, and robot training scenarios, though it notes longer simulations may experience shifts in detail and object permanence.