Hassabis, a Nobel laureate in Physics and the leader of Google DeepMind, has introduced Gemini Omni, the most all-encompassing iteration within the Gemini model family. This model is adept at seamlessly and profoundly processing multimodal information. Throughout the demonstration, Gemini Omni displayed a wide array of capabilities, such as recognizing hand-drawn fish and creating animations, transforming knowledge points into visual narratives involving marbles moving along paths, and identifying astronomical sketches accompanied by explanations. These demonstrations underscored its potential in scientific reasoning and visualization. Furthermore, Gemini Omni has been seamlessly integrated into the Gemini App, where it functions as the default intelligent engine and serves as a unified model foundation for forthcoming services.
