Why OpenAI Models Struggle with PDF Extraction(and Why Gemini Fairs Much Better)

  • Does this only apply to PDFs with many images? If it is mostly text and table, could it just extract plain text?