Comment by fragmede

Tables in PDFs still confuse traditional OCR engines. VLMs do better in some cases (though not this one, apparently).