Google announced Genie 4 on Oct. 27, a new world model that the company said brings native audio to real-time interactive simulation. Google described the announcement as research.
Google said it has started rolling out Genie 4 to a limited group of trusted testers. It said the model is coming soon to Google AI Ultra subscribers in Project Genie. Early access partners are also due to get it in the Gemini API and AI Studio, according to the company.
According to Google, users can set an event in motion and steer scenes in real time using natural language. Examples the company gave include moving the camera, changing the weather and introducing new characters. Google said every world also generates native audio on the fly, including ambient sounds, music and interactive dialogue.
The company said Genie 4 builds on capabilities from its Gemini family of AI models. Gemini Omni anchors simulations in real-world appearance and behavior, Google said. Other Gemini models provide the complex reasoning needed to help maintain a simulation’s coherence and continuity, it said.
Google said simulating the world is a crucial step toward what it called AGI, or artificial general intelligence. It also said such simulation can enable safe training environments for robots and agents. The company said it can also amplify creative tools in industries such as filmmaking, journalism, gaming and manufacturing. These are the company’s statements about the technology’s potential.
Google pointed readers to demonstrations on the Google DeepMind blog.