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

AI-generated text detection research benchmarks – Page 7 · SOTA2 Research

AI-generated text detection

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

OUTFOX Non-attacked (test)

99.1Avg Recall

5

Feb 26, 2026

Deepfake Domain-specific & Model-specific (In-distribution)

99.77Average Recall

5

Feb 26, 2026

Deepfake Cross-domains Model-specific (In-distribution)

96.73Avg Recall

5

Feb 26, 2026

Student Essays Out-of-Domain

97.7F1 Score

5

Feb 26, 2026

Creative Writing Out-of-Domain

95.7F1 Score

5

Feb 26, 2026

News Out-of-Domain

97.9F1 Score

5

Feb 26, 2026

All Domains In-Domain

99F1 Score

5

Feb 26, 2026

Student Essays In-Domain

99.5F1 Score

5

Feb 26, 2026

Creative Writing In-Domain

98.4F1 Score

5

Feb 26, 2026

99.5F1 Score

5

Feb 26, 2026

TOEFL 91 (test)

96.7Accuracy

5

Feb 26, 2026

TOEFL 11 (test)

100Accuracy

5

Feb 26, 2026

99.2Accuracy

5

Feb 26, 2026

Claude-generated (test)

92.2F1 Score

5

Feb 26, 2026

Prompt variants various (test)

99.5F1 Score

5

Feb 26, 2026

Best LLM Finetune (Phi-4)

99.6Accuracy

4

Feb 26, 2026

Mixed-source (test)

95.33OA

4

Feb 26, 2026

Cross-domain (test)

93OA

4

Feb 26, 2026

In-domain (test)

100OA

4

Feb 26, 2026

HC3 cross-benchmark transfer

99.8AUROC (Finance)

3

Jul 7, 2026

84.78TPR @ 1% FPR

3

Jun 2, 2026

MAGE DeepSeek-R1 OOD

71Accuracy

3

Apr 16, 2026

MAGE Claude-sonnet-4-5 OOD

57.1Accuracy

3

Apr 16, 2026

68.7Accuracy

3

Apr 16, 2026

61.8Accuracy

3

Apr 16, 2026

Page 7 of 8

Popular datasets

XSum, SQuAD, WritingPrompts M4GT-Bench Wikipedia MAGE RAID

Follow for updates

Join SOTA2 Contact