FAQ from Genie 2 by Google DeepMind
What is Genie 2 by Google DeepMind?
Genie 2 is a state-of-the-art generative world model that synthesizes rich, interactive 3D environments directly from textual or visual cues—enabling seamless exploration, agent embodiment, and dynamic world adaptation.
How to use Genie 2 by Google DeepMind?
Enter any descriptive prompt on the Genie 2 web interface, and the system will render a fully simulated, navigable 3D world—with physics, interactivity, and behavioral logic—all in real time.
Is Genie 2 publicly available?
Genie 2 is currently in closed research preview. Public access is expected later this year—sign up for early updates via the official website.
Can Genie 2 generate worlds from images alone?
Yes. In addition to text, Genie 2 supports single-image conditioning—turning sketches, concept art, or even smartphone photos into immersive, playable 3D spaces with contextual understanding.