While traditional Optical Character Recognition (OCR) had its share of challenges, large language models have made impressive strides in addressing them. Despite this progress, there's still room for improvement. Nonetheless, we've achieved a substantial advancement in the document parsing.