Researchers at the Beijing Academy of Artificial Intelligence have released OmniGen 2, an open-source system for text-to-image generation, image editing, and contextual image creation.