Vision Language Models Are Blind - Interesting Results From Gpt4O And Other Vlms