Expressive speech generation on EchoMind
2.76WERQwen3-Omni-30B
Evaluation results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Qwen3-Omni-30BModel Type=Industry SLM2026.04 | 2.76 | 49.76 | 25.03 | 6.23 | 7.14 | 1.41 | 7.7 | 5.62 | 4.08 | 4.28 | 4.42 | 4.25 | |
| Qwen2.5-Omni-7BModel Type=Industry SLM2026.04 | 3.75 | 47.77 | 23.1 | 6.23 | 7.08 | 1.42 | 7.79 | 5.63 | 3.62 | 3.9 | 4.09 | 3.86 | |
| CosyVoice2Model Type=TTS2026.04 | 3.8 | 51.2 | 39.07 | 6.21 | 7.09 | 1.42 | 7.67 | 5.6 | 4.15 | 4.25 | 4.34 | 4.23 | |
| Qwen2.5-Omni-3BModel Type=Industry SLM2026.04 | 4.14 | 42.88 | 21.44 | 6.15 | 6.9 | 1.47 | 7.74 | 5.57 | 3.71 | 3.84 | 3.96 | 3.79 | |
| UltraVoice-7BModel Type=Expressive SLM2026.04 | 4.21 | 49.53 | 17.34 | 6.11 | 7.13 | 1.44 | 7.55 | 5.56 | 4.01 | 4.17 | 4.26 | 4.11 | |
| SA-SLMModel Type=Expressive SLM2026.04 | 4.56 | 63.44 | 35.61 | 6.12 | 7.19 | 1.41 | 7.82 | 5.64 | 4.24 | 4.36 | 4.49 | 4.33 | |
| Step Audio2 MiniModel Type=Industry SLM2026.04 | 4.87 | 50.92 | 22.92 | 6.12 | 7.06 | 1.44 | 7.68 | 5.58 | 4 | 4.19 | 4.43 | 4.15 | |
| Kimi AudioModel Type=Industry SLM2026.04 | 6.99 | 46.98 | 22.64 | 5.31 | 5.98 | 1.49 | 6.87 | 4.91 | 3.77 | 4.02 | 4.15 | 3.97 | |
| GPT-4o-AudioModel Type=Commercial SLM2026.04 | 7.63 | 70.44 | 45.04 | 6.22 | 7.37 | 1.41 | 8 | 5.75 | 4.27 | 4.43 | 4.58 | 4.41 | |
| GLM-4-VoiceModel Type=Industry SLM2026.04 | 8.88 | 43.71 | 23.84 | 6.22 | 6.94 | 1.42 | 7.57 | 5.54 | 3.95 | 4.11 | 4.29 | 4.09 |