Over the past few years, those models have been upgraded to become omni LLMs -- basically, they are multimodal (can handle text, image and ... be able to select certain elements in an AI-generated ...