Datalab Releases OmniExtractBench to Fix Document Extraction Benchmarks

Vik Paruchuri's Datalab team released OmniExtractBench, a 620-document benchmark consolidating multiple existing datasets to correct scoring and ground-truth flaws in prior document extraction evaluations.

2026-09-17 ~ 2026-09-17 · 4 related posts

• Datalab launches OmniExtractBench: 620 docs to fix biased structured-extraction benchmarks — VikParuchuri · 2026-09-17
• Document extraction bake-off: all methods rerun same window with unified scoring, code open-sourced — VikParuchuri · 2026-09-17
• Marker author builds new document extraction benchmark, fixing major scoring flaws in existing ones — VikParuchuri · 2026-09-17
• Mistral OCR's Vik Paruchuri launches a new document-extraction benchmark — VikParuchuri · 2026-09-17