Text-to-Speech on ZTTS1-Eval In-The-Wild (fr)
2.17WEREleven Labs V3
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Eleven Labs V3supports zero-shot voice cloning=false2026.06 | 2.17 | 3.28 | 13.6 | |
| Cartesia Sonic 3.52026.06 | 2.52 | 2.69 | 75.5 | |
| Qwen 3 TTS 1.7B2026.06 | 2.62 | 2.84 | 67.1 | |
| ZONOS2 8B Quality ModeQuality Mode=true2026.06 | 2.79 | 2.74 | 61.2 | |
| Fish S2 Pro2026.06 | 3.45 | 2.67 | 68.3 | |
| Gemini 3.1 Flashsupports zero-shot voice cloning=false2026.06 | 3.56 | 3.45 | 15 | |
| Inworld TTS 22026.06 | 4.4 | 2.76 | 54.9 | |
| VoxCPM 22026.06 | 4.54 | 2.36 | 72.2 | |
| ZONOS2 8BQuality Mode=false2026.06 | 4.56 | 2.37 | 67.7 | |
| Ground truth2026.06 | 8.08 | 2.27 | 77.6 |