Google Announces Genie 4 World Model With Native Audio and Real-Time Prompting

Genie 4 is Google's first real-time world model to support native audio generation, Android Authority reported. It improves on last year's Genie 3 in several key areas and builds on Gemini Omni, which Google says anchors Genie 4's simulations in real-world appearance and behavior.

Google says Genie 4 can keep a scene consistent for up to two minutes, double Genie 3's limit. With natural-language prompts, users can shape the world in real time, changing the weather, moving the camera and introducing new people or objects while the model keeps the scene consistent, according to the report.

The announcement highlights a branch of generative AI known as world models. Such systems build an internal simulation of an environment and use physics, spatial relationships and time progression to produce consistent output. Large language models use token data to predict the next word, while world models predict how space, time, gravity and other physical forces will evolve in an environment, Android Authority reported.

Android Authority noted that generative AI is often associated with video and image generation tools such as Gemini Omni or Nano Banana, but world models are a research focus for many AI labs. Google has positioned Genie 4 as an improvement over Genie 3, with longer scene consistency and native audio, though access remains limited during the research preview.