Codesota · Models · olmOCR v0.4.0Allen AI9 results · 1 benchmarks
Model card

olmOCR v0.4.0.

Allen AIopen-sourceOCR PipelineApache 2.01 current SOTA
§ 01 · Benchmarks

Every benchmark olmOCR v0.4.0 has a recorded score for.

#BenchmarkArea · TaskMetricValueRankDateSource
01olmOCR-BenchComputer Vision · Document Parsingheaders-footers96.1%#1/4source ↗
02olmOCR-BenchComputer Vision · Document Parsingbase99.7%#2/4source ↗
03olmOCR-BenchComputer Vision · Document Parsingold-scans-math82.3%#2/4source ↗
04olmOCR-BenchComputer Vision · Document Parsingold-scans47.7%#3/5source ↗
05olmOCR-BenchComputer Vision · Document Parsingarxiv83.0%#3/5source ↗
06olmOCR-BenchComputer Vision · Document Parsinglong-tiny-text81.9%#3/4source ↗
07olmOCR-BenchComputer Vision · Document Parsingmulti-column83.7%#3/4source ↗
08olmOCR-BenchComputer Vision · Document Parsingtables84.9%#4/5source ↗
09olmOCR-BenchComputer Vision · Document Parsingpass-rate82.4%#5/21source ↗
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 02 · Strengths by area

Where olmOCR v0.4.0 actually performs.

Computer Vision
1
benchmark
avg rank #2.9 · 1 SOTA
§ 04 · Related models

Other Allen AI models scored on Codesota.

olmOCR v0.3.0
2 results
olmOCR-7B
1 result
§ 05 · Sources & freshness

Where these numbers come from.

paper
8
results
alphaxiv-leaderboard
1
result
0 of 9 rows marked verified.