Text-to-Speech on ZTTS1-Eval In-The-Wild (hi)
7.13WERVoxCPM 2
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| VoxCPM 22026.06 | 7.13 | 2.41 | 73.4 | |
| Eleven Labs V3supports zero-shot voice cloning=false2026.06 | 7.36 | 3.47 | 16.3 | |
| Cartesia Sonic 3.52026.06 | 7.87 | 2.71 | 76.1 | |
| Gemini 3.1 Flashsupports zero-shot voice cloning=false2026.06 | 8.73 | 3.6 | 18.1 | |
| ZONOS2 8B Quality ModeQuality Mode=true2026.06 | 9.04 | 2.73 | 62.3 | |
| Inworld TTS 22026.06 | 10.14 | 2.65 | 60.5 | |
| Fish S2 Pro2026.06 | 11.69 | 2.69 | 69.9 | |
| Ground truth2026.06 | 14.2 | 2.22 | 78.7 | |
| ZONOS2 8BQuality Mode=false2026.06 | 15.5 | 2.47 | 66.3 |