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

Text Summarization research benchmarks – Page 3 · SOTA2 Research

Text Summarization

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

CNN/DailyMail LLM Reference (test)

45.1ROUGE-1

4

Apr 23, 2026

20.5Hal. Rate

4

Apr 21, 2026

43.1ROUGE-2

4

Apr 14, 2026

MiniLingua-1b-Instruct

18.7Score

4

Feb 26, 2026

65.36ROUGE-1

4

Feb 26, 2026

5.6ROUGE-L

4

Feb 26, 2026

13.5ROUGE-L

4

Feb 26, 2026

12.5ROUGE-L

4

Feb 26, 2026

17.1ROUGE-L

4

Feb 26, 2026

47.3ROUGE-L

4

Feb 26, 2026

35.021st Percent

4

Feb 26, 2026

SciTLDR 76 gold papers

0.6MRR

4

Feb 26, 2026

47Rank 1 Proportion

4

Feb 26, 2026

11.29ROUGE-L

3

Feb 26, 2026

0.474Avg Spearman Corr

3

Feb 26, 2026

Vietnamese News Dataset original (val)

61.1ROUGE-1

3

Feb 26, 2026

VNDS with duplicate article removal 1.0 (test)

60.51ROUGE-1 Score

3

Feb 26, 2026

VNDS with duplicate article removal 1.0 (val)

60.55ROUGE-1

3

Feb 26, 2026

2.08Throughput

2

May 27, 2026

Pioneer Agent (Qwen3-8B)

17.9ROUGE-2

2

Apr 14, 2026

7.15ROUGE-L

2

Mar 5, 2026

14.01ROUGE-L

2

Mar 5, 2026

—

—Primary metric

0

Feb 26, 2026

Page 3 of 3

Popular datasets

CNN Dailymail Xsum CNN/DM SummEval Gigaword

Follow for updates

Join SOTA2 Contact