Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Paralinguistic-aware Response Generation on Paralinguistic-aware Gender Evaluation Set
Loading...
98.5
PA-score
GPT-4.1 baseline
4.9
29.2
53.5
77.8
Mar 12, 2026
PA-score
PA-rate
ParaS2S
Updated 1mo ago
Evaluation Results
Method
Method
Links
PA-score
PA-rate
ParaS2S
GPT-4.1 baseline
Layers=N/A, ADCH=N/A
2026.03
98.5
98.5
4
Kimi-Audio
Layers=0-14, ADCH=×
2026.03
97
97
3.99
Qwen2.5-Omni (PE-FT)
Layers=0-14, ADCH=✓
2026.03
96.5
96.5
3.99
Kimi-Audio (PE-FT)
Layers=0-14, ADCH=✓
2026.03
96.5
96.5
3.98
Qwen2.5-Omni
Layers=0-27, ADCH=×
2026.03
95.5
96
3.96
Kimi-Audio
Layers=0-27, ADCH=×
2026.03
95
95
3.98
Qwen2.5-Omni
Layers=0-14, ADCH=×
2026.03
94.5
95.5
3.93
Qwen2.5-Omni
Layers=N/A, ADCH=N/A
2026.03
10
15
3.67
Kimi-Audio
Layers=N/A, ADCH=N/A
2026.03
8.5
12.5
3.8
Feedback
Search any
task
Search any
task