Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Spoken Question Answering on UltraEval-Audio WebQ
Loading...
55.95
S2T Score (Task Correctness)
Qwen3-Omni-30B
31.25
37.6625
44.075
50.4875
Jun 29, 2026
S2T Score (Task Correctness)
S2S Score (Task Correctness)
Word Error Rate (WER)
Updated 24d ago
Evaluation Results
Method
Method
Links
S2T Score (Task Correctness)
S2S Score (Task Correctness)
Word Error Rate (WER)
Qwen3-Omni-30B
2026.06
55.95
52.51
14.92
GPT-4o
2026.06
50.98
-
-
VocalNet-8B
streaming mode=true
2026.06
44.05
39.27
7.68
Qwen2.5-Omni-7B
2026.06
42.18
39.42
21.5
Backbone-LLM-7B
2026.06
42.18
-
-
PRIME-Speech-9B
2026.06
42.04
40.18
3.33
Kimi-Audio-7B
2026.06
41.98
35.37
14.85
Step-Audio-2-Mini-7B
2026.06
33.02
31.69
8.56
GLM-4-Voice-9B
2026.06
32.2
15.9
-
Feedback
Search any
task
Search any
task