Comment by faebi

Did you try other vision models such as ChatGPT and Grok? I'm doing something similar but struggled to find good comparisons in between the vision models in terms OCR and document understanding.