Details
- Google DeepMind announced Genie 4, a world model designed for real-time interactive simulation.
- Users can steer generated environments with natural-language commands, including changing weather, moving the camera, or adding people and objects.
- The system generates synchronized visuals and audio, including music, ambient sounds, and interactive dialogue.
- Genie 4 combines Gemini-family capabilities to anchor simulations in realistic appearance and behavior while maintaining scene coherence and continuity.
- Google is testing the model with trusted partners in filmmaking, game design, education, and robotic simulation, including visual-effects pre-visualization.
- Safety guardrails are built into the system, and generated audio and video carry Google’s SynthID watermark.
- The research preview is rolling out to trusted testers, with planned access for Google AI Ultra subscribers through Project Genie and selected developers through the Gemini API and Google AI Studio.
- Google has not announced general availability or developer pricing.
Impact
Genie 4 pushes Google’s world-model strategy toward practical creative and simulation workflows by combining real-time scene control with native audio. Its phased release limits immediate adoption, but access through Project Genie, the Gemini API, and AI Studio could let Google test both consumer and developer demand before broader deployment. Watermarking and safety controls also position the launch within growing pressure for traceability in synthetic media.