Text-to-Speech on LibriSpeech clean (WER)
3.46WERUniAudio 2.0
Evaluation results
| Method | Links | |
|---|---|---|
| UniAudio 2.0evaluation_setting=Seen-task evaluation2026.02 | 3.46 | |
| Qwen2.5-Omni 7Bevaluation_setting=Seen-task evaluation2026.02 | 4.28 | |
| MiMo-Audio-7B-Instructevaluation_setting=Seen-task evaluation2026.02 | 4.74 | |
| Phi-4 Multimodal2026.01 | 4.8 | |
| MoST2026.01 | 6 | |
| SeamlessM4T-v22026.01 | 6.3 | |
| SpiritLM2026.01 | 6.7 | |
| MinMo2026.01 | 6.7 | |
| Moshi2026.01 | 7 | |
| AudioLM2026.01 | 9.2 | |
| LLaMA-Omni22026.01 | 10.1 | |
| SpeechGPT2026.01 | 14.1 |