Text-to-Speech on ZTTS1-Eval In-The-Wild (pt)
2.06WEREleven Labs V3
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Eleven Labs V3supports zero-shot voice cloning=false2026.06 | 2.06 | 3.39 | 22.1 | |
| Cartesia Sonic 3.52026.06 | 2.1 | 2.68 | 77.6 | |
| Qwen 3 TTS 1.7B2026.06 | 2.2 | 2.85 | 69.4 | |
| Fish S2 Pro2026.06 | 2.86 | 2.57 | 71.8 | |
| ZONOS2 8B Quality ModeQuality Mode=true2026.06 | 2.94 | 2.75 | 63.8 | |
| Gemini 3.1 Flashsupports zero-shot voice cloning=false2026.06 | 3.21 | 3.43 | 24.1 | |
| Inworld TTS 22026.06 | 3.87 | 2.7 | 61.4 | |
| VoxCPM 22026.06 | 4.09 | 2.32 | 73.5 | |
| ZONOS2 8BQuality Mode=false2026.06 | 6.73 | 2.33 | 70.4 | |
| Ground truth2026.06 | 7.56 | 2.23 | 79.5 |