Content Moderation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
83.1Average F1 Score
30
Apr 7, 2026
95.76Accuracy
27
Mar 11, 2026
86.72Overall Composite Score
22
Jun 29, 2026
86.1F1 Score
21
Apr 7, 2026
89.7Politics Accuracy
15
Apr 14, 2026
95.2Sexual Accuracy
14
May 19, 2026
95.4Score (%)
12
Jun 25, 2026
89.2Average F1
11
Apr 14, 2026
81.4Accuracy
8
Mar 18, 2026
83Precision
8
Feb 26, 2026
94Precision
8
Feb 26, 2026
94Precision
8
Feb 26, 2026
89.2AUC-ROC
7
Feb 26, 2026
0.89AUROC
6
May 13, 2026
86.1Accuracy
4
Mar 17, 2026
3.6Refusal Rate (Prod)
4
Feb 26, 2026
72Hate
4
Feb 26, 2026
64.7Abusive Rate
4
Feb 26, 2026
99.38Ad Recall
3
Apr 1, 2026
10LLM Calls
3
Mar 4, 2026
9LLM Calls
3
Mar 4, 2026
9LLM Calls
3
Mar 4, 2026
6LLM Calls
3
Mar 4, 2026
7LLM Calls
3
Mar 4, 2026
6LLM Calls
3
Mar 4, 2026