- dataset:
    id: allenai/olmOCR-bench
    task_id: old_scans
  value: 37.8
  date: "2026-06-27"
  source:
    url: https://github.com/davanstrien/ocr-bench/blob/99f7550c/experiments/olmocr-bench-oldscans/BENCHMARKING.md
    name: ocr-bench — old_scans multi-model comparison
    user: davanstrien
  notes: "old_scans.jsonl sub-score (present/absent/order); markdown mode, non-thinking, greedy, 170 DPI. NuExtract3 leads the field on present (41.6) — see source for the full sub-score breakdown."