Text-to-Speech on ZTTS1-Eval In-The-Wild (en)
1.05WERQwen 3 TTS 1.7B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen 3 TTS 1.7B2026.06 | 1.05 | 3.2 | 61.5 | |
| Eleven Labs V3supports zero-shot voice cloning=false2026.06 | 1.35 | 3.61 | 6.3 | |
| Cartesia Sonic 3.52026.06 | 1.4 | 3.05 | 70.2 | |
| VoxCPM 22026.06 | 1.69 | 2.51 | 68.1 | |
| Inworld TTS 22026.06 | 1.89 | 3.02 | 53 | |
| Fish S2 Pro2026.06 | 2.09 | 2.92 | 65 | |
| Gemini 3.1 Flashsupports zero-shot voice cloning=false2026.06 | 2.12 | 3.78 | 10.2 | |
| ZONOS2 8B Quality ModeQuality Mode=true2026.06 | 2.21 | 2.99 | 56.9 | |
| ZONOS2 8BQuality Mode=false2026.06 | 4.7 | 2.44 | 67 | |
| Ground truth2026.06 | 5.61 | 2.48 | 75.9 |