Vision LLMs vs OCR benchmark on document QA

View on Reddit

A user benchmarked vision-capable LLMs against OCR pipelines on 30 image-heavy PDFs from MMLongBench-Doc with 171 questions. The comparison evaluates the 'just attach the PDF' pattern vs traditional OCR for long-document QA.