This past summer, Google DeepMind Genie 3 starts. This is known as the virtual world, an AI system capable of generating images and reacting as the user moves around the environment simulated by the software. At the time, DeepMind pitched Genie 3 as a tool for training AI agents. Now, it’s making the model available for people outside of Google to test Project Genius.

To get started, you need Google’s $250 per month AI Ultra plan see Project Genie. You must also live in the US and be 18 years or older. At launch, Project Genie offers three different modes of interaction: World Sketching, exploration and remixing. The first saw Google Nano Banana Pro model that creates the source image that Genie 3 will use to create the world you’ll explore later. In this stage, you can describe your character, define the camera perspective – first person, third person or isometric – and how you want to explore the world that Genie 3 is about to create. Before you jump into creating the model, Nano Banana Pro will “sketch” what you see so you can make tweaks. It is also possible to write your own prompts for the worlds that others use Genie to create.

One thing to keep in mind is that the Genie 3 is not a gaming machine. While its outputs may look game-like, and it may simulate physical interactions, there are no traditional game mechanics here. Generations are also limited to 60 seconds, as is the presentation, which is limited to 24 frames per second and 720p. However, if you are an AI Ultra subscriber, this is a cool opportunity to see the bleeding edge of what DeepMind has been working on for the past two years.