Text-to-Speech on ZTTS1-Eval In-The-Wild (de)
2.11WERQwen 3 TTS 1.7B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen 3 TTS 1.7B2026.06 | 2.11 | 3.12 | 68.3 | |
| Gemini 3.1 Flashsupports zero-shot voice cloning=false2026.06 | 2.4 | 3.67 | 17.5 | |
| Eleven Labs V3supports zero-shot voice cloning=false2026.06 | 2.5 | 3.57 | 15.3 | |
| Fish S2 Pro2026.06 | 2.65 | 2.9 | 69.4 | |
| VoxCPM 22026.06 | 2.84 | 2.6 | 72.8 | |
| Cartesia Sonic 3.52026.06 | 3 | 2.98 | 76.7 | |
| ZONOS2 8B Quality ModeQuality Mode=true2026.06 | 3.37 | 2.92 | 63.1 | |
| Inworld TTS 22026.06 | 3.54 | 3.03 | 58.7 | |
| ZONOS2 8BQuality Mode=false2026.06 | 5.84 | 2.52 | 69.4 | |
| Ground truth2026.06 | 6.77 | 2.52 | 78.5 |