Sora is a cutting-edge generative model capable of creating high fidelity videos by training on a vast array of video and image data. It leverages a novel approach using spacetime patches and operates as a diffusion transformer, enabling the generation of diverse video content. This model represents a promising step towards developing general-purpose simulators of the physical and digital world, capable of simulating intricate aspects of reality.

Basic information

Country

USA

Group Name

OpenAI

Source

Source