Google Genie

// about Google Genie

Genie is Google DeepMind's world model, which turns a text or image prompt into an interactive environment you can move around in. Genie 4, previewed on October 27, 2026, is the first DeepMind world model to generate audio natively. You can change the weather, move the camera, or add characters mid-run without restarting. Scenes stay consistent for about two minutes, and Google says output carries a SynthID watermark. It's a research preview for trusted testers, with Google AI Ultra access in Project Genie and Gemini API early access planned in the coming weeks.