Ranking of LLMs for agentic tasks
Generate descriptions from images and text prompts
Manipulate images using drag points