Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Document Parsing research benchmarks – Page 2 · SOTA2 Research

Document Parsing

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

medical invoice (test)

0.279FMR

10

Feb 26, 2026

pdf-parse bench

8.26Edit Metric

9

Mar 16, 2026

OmniDocBench v1.5

PaddleOCR-VL-1.5

944.4Total Time (s)

9

Feb 26, 2026

Qwen3.5-397B-A17B

81.85Overall Score

8

Mar 4, 2026

XDocParse (test)

1,323.7Elo Rating

7

Mar 16, 2026

OmniDocBench v1.5 (test)

1,128Elo Rating

7

Mar 16, 2026

olmOCR-Bench (test)

1,180.4Elo Rating

7

Mar 16, 2026

96.5Score

7

Mar 12, 2026

94Overall Score

7

Mar 12, 2026

96.49TEDS

7

Mar 4, 2026

86Score

6

Mar 12, 2026

OmniDocBench v1.5

1.96AAL

6

Feb 27, 2026

86.31Overall Accuracy

4

Jun 24, 2026

Multilingual Document Parsing Dataset

39Performance (Arabic)

4

May 28, 2026

Multilingual OCR Japanese

PaddleOCR-VL-0.9B

0.086Edit Distance

4

Mar 26, 2026

Latin Multilingual

PaddleOCR-VL-0.9B

0.013Edit Distance

4

Mar 26, 2026

PaddleOCR-VL-0.9B

0.109Edit Distance

4

Mar 26, 2026

PaddleOCR-VL-0.9B

0.097Edit Distance

4

Mar 26, 2026

PaddleOCR-VL-0.9B

0.011Edit Distance

4

Mar 26, 2026

Thai Multilingual OCR

PaddleOCR-VL-0.9B

0.081Edit Distance

4

Mar 26, 2026

Multilingual OCR Greek

PaddleOCR-VL-0.9B

0.135Edit Distance

4

Mar 26, 2026

PaddleOCR-VL-0.9B

0.043Edit Distance

4

Mar 26, 2026

Multilingual OCR Korean

PaddleOCR-VL-0.9B

5.2Edit Distance

4

Mar 26, 2026

Arabic Multilingual OCR

PaddleOCR-VL-0.9B

0.122Edit Distance

4

Mar 26, 2026

73.7ANLS

4

Mar 5, 2026

Page 2 of 3

Popular datasets

OmniDocBench Multilingual OCR DocPTBench Ocean-OCR-Bench olmOCR-Bench

Follow for updates

Join SOTA2 Contact