Text-to-Music Generation
Benchmarks
DatasetSOTA methodMetricTrendResultsLast updated
2.18FAD
20
Feb 26, 2026
1.01KLD
19
Jun 12, 2026
0.417FAD
14
May 21, 2026
0.531ms-CLAP Score
13
May 21, 2026
2FAD
7
Feb 26, 2026
85.7T2M-QLT
6
Feb 26, 2026
5.18REL (General)
6
Feb 26, 2026
41.08Overall Preference Score
5
Feb 26, 2026
72.17FD_openl3
5
Feb 26, 2026
4OVL
4
Feb 26, 2026
4.2OVL (Overall Likeness)
4
Feb 26, 2026
0.518MuLan Similarity (Audio vs GT Text)
4
Feb 26, 2026
56.3MuLan Similarity (Audio to GT Text)
4
Feb 26, 2026
0.33CLAP Similarity (Benign, User Question)
3
Jun 1, 2026
55.5Musicality
2
Feb 26, 2026
0.541Musicality
2
Feb 26, 2026
4.09REL (General Audience)
1
Feb 26, 2026
—
—Primary metric
0
Feb 26, 2026