List your company

Zero-shot Text-to-Speech on Seed-TTS en (test)

1.08WERQwen3-TTS
0.623.7256.839.935May 7, 2026May 11, 2026May 15, 2026May 19, 2026May 23, 2026May 27, 2026Jun 1, 2026

Evaluation results

MethodLinks
Qwen3-TTSRTF=1.7542026.051.0871—
X-Voices2RTF=0.0732026.051.365—
Fish Audio S2RTF=4.8012026.051.3765—
Qwen3-TTSSize=1.7B, Emilia=×2026.051.43471.24.178
OmniVoiceSize=0.6B, Emilia=×2026.051.47574.13.896
LEMAS-TTSRTF=0.1312026.051.4962—
X-Voices1RTF=0.0732026.051.5365—
OmniVoiceRTF=0.1982026.051.674—
Qwen3-TTSSize=0.6B, Emilia=×2026.051.71170.64.155
IndexTTS2Size=1.5B, Emilia=INCL.2026.051.73570.73.651
GibbsTTSSize=0.4B, Emilia=✓2026.051.77774.33.651
F5-TTSRTF=0.0652026.051.8966—
MOSS-TTSRTF=0.6432026.051.9269—
Ground truthSize=—, Emilia=—2026.052.0273.43.527
F5-TTSSize=0.3B, Emilia=✓2026.052.0265.43.762
CosyVoice 3Size=0.5B, Emilia=×2026.052.14569.63.949
F5-TTSmode=zero-shot2026.062.1575.53.68
FireRedTTS-2Size=1.5B, Emilia=×2026.052.18766.43.697
Spark-TTSSize=0.5B, Emilia=✓2026.052.22157.23.947
Codec reconstructedSize=—, Emilia=—2026.052.22969.53.407
CosyVoice 2Size=0.5B, Emilia=×2026.052.50665.64.164
UniVocalmode=zero-shot2026.062.6970.34.21
CosyVoice 2mode=zero-shot2026.062.9674.44.18
MaskGCTSize=1.5B, Emilia=✓2026.053.76371.63.582
Vevo 1.5mode=zero-shot2026.0612.5871.83.68