Benchmark

olmOCR-Bench

Measures: 1,403 PDFs with 7,010 machine-checkable unit tests (text present or absent, reading order, table cells, math formulas) across arXiv math, old scans, tables, headers and footers, multi-column and tiny text; score = % of tests passed.

Maintainer
Allen Institute for AI (Ai2)
Introduced
2025
Status
Active
Saturation
Headroom
Typical provenance
Mixed
In scoring set
Yes

PaperOfficial leaderboard

Leaderboard

Independent and third-party results only.

No independent or third-party results yet. Show all results

olmOCR-Bench: leaderboard and what it measures · supermarket.dev