1,121
±14 95% CI
67%
Win rate
2,387
Battles
AudioLab Leaderboards
Pairwise preference rankings and independent benchmarks for leading speech models.
Valid votes
25.5K
Models
22
Method
Bradley–Terry MLE
Text to Speech rankings
Models are ordered by rating. Models are marked preliminary until 200 valid battles; disabled models are hidden. Newly added models appear unranked at the bottom until they record their first battles.
1,109
±14 95% CI
66%
Win rate
2,372
Battles
1,085
±13 95% CI
62%
Win rate
2,353
Battles
1,057
±14 95% CI
58%
Win rate
2,312
Battles
1,054
±14 95% CI
58%
Win rate
2,410
Battles
1,037
±15 95% CI
56%
Win rate
2,443
Battles
1,032
±17 95% CI
54%
Win rate
1,701
Battles
1,032
±14 95% CI
54%
Win rate
2,348
Battles
1,027
±13 95% CI
54%
Win rate
2,464
Battles
1,025
±15 95% CI
54%
Win rate
1,998
Battles
1,021
±13 95% CI
53%
Win rate
2,417
Battles
1,009
±15 95% CI
51%
Win rate
2,334
Battles
991
±14 95% CI
49%
Win rate
2,336
Battles
985
±15 95% CI
47%
Win rate
2,409
Battles
966
±16 95% CI
45%
Win rate
2,387
Battles
949
±13 95% CI
43%
Win rate
2,414
Battles
944
±13 95% CI
42%
Win rate
2,422
Battles
943
±16 95% CI
42%
Win rate
2,384
Battles
938
±15 95% CI
41%
Win rate
2,365
Battles
931
±14 95% CI
41%
Win rate
2,152
Battles
895
±14 95% CI
35%
Win rate
2,468
Battles
848
±16 95% CI
29%
Win rate
2,192
Battles