A preprint research paper has introduced a new benchmark called SUPERGLASSES that evaluates how well vision-language models ...