Text-to-Speech on ZTTS1-Eval In-The-Wild (tl)
15.7WERVoxCPM 2
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| VoxCPM 22026.06 | 15.7 | 2.51 | 73 | |
| ZONOS2 8B Quality ModeQuality Mode=true2026.06 | 16.05 | 2.88 | 65 | |
| Eleven Labs V3supports zero-shot voice cloning=false2026.06 | 17.76 | 3.51 | 13.1 | |
| Qwen 3 TTS 1.7B2026.06 | 18.04 | 3.13 | 60.5 | |
| Fish S2 Pro2026.06 | 19.13 | 2.89 | 64.9 | |
| Inworld TTS 22026.06 | 19.56 | 2.99 | 59.5 | |
| ZONOS2 8BQuality Mode=false2026.06 | 23.49 | 2.52 | 68.5 | |
| Ground truth2026.06 | 25.75 | 2.24 | 78.9 |