Google’s Genie 4 world model adds audio, more realistic worlds, and real-time steering
techcrunch ·
Google DeepMind has released Genie 4, its latest world model, in a research preview that introduces photorealistic environments with real-time audio generation and interactive steering capabilities. Unlike its predecessor, Genie 4 allows users to modify scenes mid-simulation by typing prompts to change weather, camera angles, or add objects while maintaining consistency for up to two minutes. The model features improved physics and more human-like characters, alongside integrated audio that includes music, ambient noise, and dialogue. Genie 4 is closely linked with Google’s Gemini AI, which acts as a director to plan steps and check frames, enabling potential applications in robotics simulation and testing. While commercial use cases include gaming and education, early feedback highlights prototyping films as a promising application due to the model's ability to maintain creative flow. Despite these advances, Google acknowledges current limitations such as object drift during long simulations and challenges with audio accuracy and text legibility. The model is currently available to trusted testers before rolling out to Google AI Ultra subscribers and API partners in the coming weeks.
techcrunch