Genie 2 is a Google AI model for generating 2D and 3D scenes from text prompts, designed for game developers and digital content creators.
What Genie 2 does
Genie 2 helps you prototype virtual worlds and game scenes with AI-generated visuals and interactions.
- Generates 2D and 3D scenes across genres such as RPGs, racing, shooters, and platformers
- Adds physics and visual effects like reflections, smoke, gravity, and water interactions
- Creates interactive NPCs and characters that can respond to commands and perform actions
- Supports text-first creation, with optional inputs like photos or sketches
Genie 2 is currently available in English only. Examples can be found on Google DeepMind’s platform.
How to use Genie 2
The tool is aimed at speeding up content creation and early-stage prototyping.
- Open Genie 2 in a browser or install the app (if available)
- Enter a text prompt or upload reference visuals (e.g., sketches)
- Adjust generation settings such as genre, scale, or lighting
- Run generation to produce a scene, world, or object
- Save the result or integrate it into your project
Notes and limitations
- Supports both simple and complex models
- Generation may be limited by time and detail constraints
- Requires significant compute resources
- Still in development; currently limited to testers, with no public release date announced
