List your company
ResearchFormerly WizWand

Speech Recognition on Common Voice EN

6.76WERStep-Audio2-Mini
4.548419.476734.40549.3333Dec 23, 2025Jan 15, 2026Feb 8, 2026Mar 4, 2026Mar 28, 2026Apr 21, 2026May 15, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Step-Audio2-Mini2025.126.76
Audio-FlamingoModel version=32025.127.4
Qwen-TTS-Tokenizer-25Hz (Stage 1)Codebook Size=32768, FPS=252026.017.51
Fun-Audio-ChatModel size=30B-A3B2025.127.79
Fun-Audio-ChatModel size=8B2025.128.88
Kimi-Audio2025.1210.31
Qwen-TTS-Tokenizer-25Hz (Stage 2)Codebook Size=32768, FPS=252026.0110.4
S3 Tokenizer(FSQ)Codebook Size=6561, FPS=252026.0110.67
SymphonySource category=Corti, Real-time processing=true2026.0511.2
ElevenLabsSource category=Closed-source, Real-time processing=true2026.0511.3
S3 Tokenizer(VQ)Codebook Size=4096, FPS=252026.0111.56
ParakeetSource category=Open-source, Real-time processing=true2026.0511.7
S3 Tokenizer(VQ)Codebook Size=4096, FPS=502026.0112.06
WhisperSource category=Open-source, Real-time processing=true2026.0513.1
OpenAISource category=Closed-source, Real-time processing=true2026.0513.3
MiMo-Audio2025.1262.05