Text-to-Speech on Seed-TTS Seed-ZH (Evaluation)
0.89CERGLM-TTS
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GLM-TTSBitrate Control=false, Architecture=Cascade (AR+NAR / NAR+NAR)2026.02 | 0.89 | 76.4 | |
| VoxCPMBitrate Control=false, Architecture=NAR / Continuous AR2026.02 | 0.93 | 77.2 | |
| IndexTTS2Bitrate Control=false, Architecture=Cascade (AR+NAR / NAR+NAR)2026.02 | 1.03 | 76.5 | |
| CosyVoice3-1.5BBitrate Control=false, Architecture=Cascade (AR+NAR / NAR+NAR)2026.02 | 1.12 | 78.1 | |
| FireRedTTS2Bitrate Control=false, Architecture=Discrete AR2026.02 | 1.14 | 73.6 | |
| VibeVoiceBitrate Control=false, Architecture=NAR / Continuous AR2026.02 | 1.16 | 74.4 | |
| OpenAudio-s1-miniBitrate Control=false, Architecture=Discrete AR2026.02 | 1.18 | 68.5 | |
| SparkTTSBitrate Control=false, Architecture=Discrete AR2026.02 | 1.2 | 67.2 | |
| CAT-TTSBitrate Control=true, Architecture=Discrete AR2026.02 | 1.23 | 78.5 | |
| CosyVoice2Bitrate Control=false, Architecture=Cascade (AR+NAR / NAR+NAR)2026.02 | 1.38 | 75.7 | |
| HiggsAudio-v2Bitrate Control=false, Architecture=Discrete AR2026.02 | 1.5 | 74 | |
| FireRedTTSBitrate Control=false, Architecture=Cascade (AR+NAR / NAR+NAR)2026.02 | 1.51 | 63.5 | |
| F5-TTSBitrate Control=false, Architecture=NAR / Continuous AR2026.02 | 1.53 | 76 | |
| LlasaBitrate Control=false, Architecture=Discrete AR2026.02 | 1.59 | 68.4 | |
| Qwen2.5-OmniBitrate Control=false, Architecture=Cascade (AR+NAR / NAR+NAR)2026.02 | 1.7 | 75.2 | |
| MaskGCTBitrate Control=false, Architecture=Cascade (AR+NAR / NAR+NAR)2026.02 | 2.27 | 77.4 |