Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Response Consistency Evaluation on TPI-Bench Shared utterances between TPI and Janus (test)
Loading...
94
BLEU
Kimi-Audio-Instruct-7B
8.72
30.86
53
75.14
Apr 19, 2026
BLEU
ROUGE-L
Updated 1mo ago
Evaluation Results
Method
Method
Links
BLEU
ROUGE-L
Kimi-Audio-Instruct-7B
Data C (Trained on cor...
2026.04
94
99
ChatGPT-4o-audio
Data C (Trained on cor...
2026.04
89
93
TPI-Base
Data C (Trained on cor...
2026.04
46
63
VITA-Audio-Instruct-7B
Data C (Trained on cor...
2026.04
42
71
TPI-VA
Data C (Trained on cor...
2026.04
39
58
Qwen2.5-Omni-7B
Data C (Trained on cor...
2026.04
31
53
Qwen3-Omni-30B-A3B-Instruct
Data C (Trained on cor...
2026.04
20
39
TPI-Full
Data C (Trained on cor...
2026.04
12
34
Feedback
Search any
task
Search any
task