| SEED-TTS EN (test) | Fish Audio S2 | WER0.99 | | 53 | 1mo ago |
| SEED-TTS-Eval ZH (test) | Fish Audio S2 | CER0.54 | | 37 | 1mo ago |
| Seed-TTS-Eval ZH-Hard (test) | | CER5.71 | | 18 | 1mo ago |
| CV3-Eval multilingual voice cloning (test) | Fish Audio S2 | WER (zh)2.65 | | 18 | 4mo ago |
| LibriHeavy-HQ | LM-Backbone LoRA for TTS | SNR40.778 | | 9 | 4mo ago |
| HiFi TTS | LM-Backbone LoRA for TTS | SNR55.854 | | 9 | 4mo ago |
| LibriSpeech-PC clean (test) | CosyVoice 2 | WER2.05 | | 8 | 1mo ago |
| Common Voice English | Chroma 1.0 | SIM Score0.81 | | 7 | 4mo ago |
| Subjective Evaluation Dataset | Qwen3-TTS | N-MOS4.8 | | 5 | 1mo ago |
| Thai Voice Cloning Short: 1-15s | JaiTTS-v1.0 | CER (%)1.94 | | 5 | 2mo ago |
| blindset-4 French (test) | Qwen3-TTS | WER0.05 | | 5 | 2mo ago |
| Thai Voice Cloning evaluation set | JaiTTS-v1.0 | RTF0.1136 | | 4 | 2mo ago |
| Thai Voice Cloning Long: 16-30s | | CER2.47 | | 4 | 2mo ago |
| blindset Chinese 4 (test) | Qwen3-TTS | CER0.09 | | 4 | 2mo ago |
| blindset-4 Arabic (test) | VoxCPM2 | WER0.209 | | 4 | 2mo ago |
| voices (unseen) | minimind-3o-moe | Similarity (CAM++)0.5702 | | 3 | 2mo ago |
| voices.pt (seen) | minimind-3o | Similarity (CAM++)64.72 | | 3 | 2mo ago |
| Voice-cloning (overall) | minimind-3o | Similarity (CAM++)59.95 | | 2 | 2mo ago |
| LRS3 (test) | | MOS4.15 | | 2 | 4mo ago |
| Voice cloning speakers (test) | - | - | | 0 | 4mo ago |