User Challenges Datalab: Benchmark Against LlamaIndex and Baidu OCR, Not Dead Textract
GhostOfTuring · x · 2026-10-28
@GhostOfTuring pushes back on Marker creator VikParuchuri's benchmark choices: Textract is dead, so test against real competition like LlamaIndex agents or Baidu Unlimited OCR. The quoted tweet gives context on Datalab's lite mode: System One OCR at $0.50 per 1k pages, claimed 10-30x cheaper than Azure/Google/AWS, 2x faster than Azure on long docs, with similar accuracy.
Related event: Datalab launches lite OCR at a fraction of cloud rivals' cost(4 posts)→