According to CoinMeta, Google has released a new generation of world model called Genie 4, which allows users to generate virtual worlds that can be explored by simply typing text. The system supports keyboard control for movement, the ability to change the weather at any time, and the addition of characters. Genie 4 can simulate collisions, aerodynamics, and character movements, and for the first time, it can generate environmental sounds, music, and real-time conversations natively. Characters will react based on the scene and user input. Genie 4 utilizes Gemini Omni to enhance the realism of graphics and behaviors, while Gemini 3.8 and Flash act as real-time commanders, continuously observing the scene and breaking down tasks. Genie 4 is currently available only to a small number of invited testers, but in the future, it will gradually be made available to Google AI Ultra, Gemini API, and AI Studio. Google hopes to use Genie 4 to create game prototypes, preview movie scenes, and enable robots and AI and agent to practice complex tasks in the virtual world.