Harmful Meme Detection
Benchmarks
DatasetSOTA methodMetricTrendResultsLast updated
76.2Accuracy
51
Jun 1, 2026
84.86Accuracy
51
Jun 1, 2026
75.79Macro-F1
49
May 29, 2026
85.5Accuracy
33
May 6, 2026
96.7Accuracy
30
May 26, 2026
83.82Accuracy
27
May 6, 2026
79.21Accuracy
16
Feb 26, 2026
79.76Precision
14
Feb 26, 2026
86.25Accuracy
11
Apr 2, 2026
88.5Racism F1
11
Feb 26, 2026
90.73Accuracy
10
Feb 26, 2026
87Accuracy
10
Feb 26, 2026
86.7F1 Score
8
Feb 26, 2026
82.4F1 Score
8
Feb 26, 2026
82.4F1 Score
8
Feb 26, 2026
87.1Racism F1
7
Feb 26, 2026