LLM-generated text detection
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
95.66LLaMA3 Score
26
Feb 26, 2026
99.22Macro F1 (All Lengths)
19
Mar 31, 2026
91.1Average Detection Score
18
Apr 3, 2026
96TPR @ FPR 1%
18
Feb 26, 2026
97.3TPR @ FPR 1%
18
Feb 26, 2026
62.7TPR @ FPR 1%
18
Feb 26, 2026
98.7TPR @ FPR 1%
18
Feb 26, 2026
99.3TPR @ FPR 1%
18
Feb 26, 2026
67.3TPR at FPR 1%
18
Feb 26, 2026
82TPR @ FPR 1%
18
Feb 26, 2026
79.3TPR @ FPR 1%
18
Feb 26, 2026
70TPR at FPR 1%
18
Feb 26, 2026
97.3TPR at FPR 1%
18
Feb 26, 2026
96.7TPR @ FPR 1%
18
Feb 26, 2026
63.3TPR @ FPR 1%
18
Feb 26, 2026
90.6Average Performance Score
16
Apr 3, 2026
99.46GPT2-XL
15
May 12, 2026
81.33AUROC
15
May 12, 2026
80.25AUROC
15
May 12, 2026
75.14AUROC
15
May 12, 2026
89.8F1 (Original)
12
Apr 30, 2026
89.2F1 (Original)
12
Apr 30, 2026
97.97AUROC (Multi-Domain)
12
May 11, 2026
94.41Avg TPR @ 5% FPR
12
Apr 7, 2026
78.99AUC-ROC (Solar)
12
Feb 26, 2026