Google DeepMind has released Genie 4, the latest version of its world model. According to the announcement, the update introduces real-time audio, more realistic simulated environments, and the ability to steer a simulation mid-run through text prompts. The model is available in a research preview limited to trusted testers.
Why it matters
Adding audio and real-time steering to a world model expands what such systems can simulate and how users interact with them during a session. These capabilities point toward more dynamic and responsive generated environments.
Who should care
Researchers and developers working with world models, simulation, and generative media may be interested in the new capabilities, though access is currently restricted to trusted testers.