Google DeepMind, the tech giant’s primary AI research and development laboratory, has played a major part in the development of Genie, a world model that’s capable of generating interactive virtual worlds.

The original Genie model launched back in March 2024 and was initially powerful enough to generate two-dimensional environments. Then in December 2024, Genie 2 came out and added the ability to generate three-dimensional worlds. Around August 2025, Genie 3 made its debut and gave users the ability to generate and modify both types of interactive worlds in real time.

This month marks the launch of Genie 4 in research preview mode, which is promised by Google to be able to generate the most photorealistic and interactive worlds to date. What makes it stand out from its predecessors is how it lets users type out prompts mid-simulation to change the scene. Also, Genie 4 can generate real-time audio that matches and syncs up with the simulated world that’s been created, which includes music, ambient sounds and interactive dialogue.

With Genie 4’s expanded capabilities tied to its interactive worlds, it could have a major impact on the way video games are created by its talented developer community.

Google has already explained how game developers have been utilizing Genie 4:

“Game developers and creative technologists are testing Genie 4 to go from an initial idea to an explorable prototype in seconds, testing visual styles, material aesthetics, and interactive mechanics through natural language prompts before writing a single line of engine code.”

Genie 4’s biggest promise is giving users the ability to generate the type of AI-generated worlds that they can actively experience, which could go a long way toward game developers creating instant mockups of what they're envisioning.

Source link