# Text-to-Speech on Seed-TTS Seed-EN (test) Source: [Text-to-Speech on Seed-TTS Seed-EN (test)]() Collection index: [llms.txt]() Data updated: 2026-05-29T10:33:07.769Z Evaluation of Text-to-Speech synthesis quality on the test split of the Seed-TTS Seed-EN dataset, assessed using Word Error Rate (WER) and Structural Similarity Index Measure (SSIM). A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Text-to-Speech | | Dataset (including any stated split) | Seed-TTS Seed-EN (test) | | Metric | WER | | Unit | PERCENT | | Direction | Lower is better | | Results loaded | 32 of 32 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [WER](): Lower is better; unit: PERCENT. - [SSIM](): Higher is better; unit: PERCENT. ## Recorded results | Method | WER | Result date | Report paper | | --- | --- | --- | --- | | ARCHI-TTS | 0.0147 | Not recorded | 69857574c97a0b3e3b5d5b84 | | DiTAR | 0.0169 | Not recorded | 69857574c97a0b3e3b5d5b84 | | SemaVoice | 0.0171 | Not recorded | 6a0beeb9dbc836b87d82c596 | | Seed-TTS\_DiT | 0.01733 | Not recorded | 69857574c97a0b3e3b5d5b84 | | F5-TTS | 0.0183 | Not recorded | 69857574c97a0b3e3b5d5b84 | | VoxCPM | 0.0185 | Not recorded | 6a0beeb9dbc836b87d82c596 | | SemaVoice-Emilia | 0.0191 | Not recorded | 6a0beeb9dbc836b87d82c596 | | OpenAudio-s1 | 0.0194 | Not recorded | 6a0beeb9dbc836b87d82c596 | | FireRedTTS-2 | 0.0195 | Not recorded | 6a0beeb9dbc836b87d82c596 | | SparkTTS | 0.0198 | Not recorded | 6a0beeb9dbc836b87d82c596 | | F5-TTS | 0.02 | Not recorded | 6a0beeb9dbc836b87d82c596 | | Ground Truth | 0.0206 | Not recorded | 69857574c97a0b3e3b5d5b84 | | VAE Resynthesized | 0.0206 | Not recorded | 69857574c97a0b3e3b5d5b84 | | Vocos Resynthesized | 0.0209 | Not recorded | 69857574c97a0b3e3b5d5b84 | | Ground Truth | 0.0214 | Not recorded | 6a0beeb9dbc836b87d82c596 | | E2 TTS | 0.0219 | Not recorded | 69857574c97a0b3e3b5d5b84 | | IndexTTS 2 | 0.0223 | Not recorded | 6a0beeb9dbc836b87d82c596 | | VoxCPM-Emilia | 0.0234 | Not recorded | 6a0beeb9dbc836b87d82c596 | | HiggsAudio-v2 | 0.0244 | Not recorded | 6a0beeb9dbc836b87d82c596 | | CosyVoice 2 | 0.0257 | Not recorded | 69857574c97a0b3e3b5d5b84 | | CosyVoice2 | 0.0257 | Not recorded | 6a0beeb9dbc836b87d82c596 | | MaskGCT | 0.0262 | Not recorded | 6a0beeb9dbc836b87d82c596 | | MaskGCT | 0.02623 | Not recorded | 69857574c97a0b3e3b5d5b84 | | Qwen2.5-Omni | 0.0272 | Not recorded | 6a0beeb9dbc836b87d82c596 | | VibeVoice | 0.0304 | Not recorded | 6a0beeb9dbc836b87d82c596 | | FireRedTTS | 0.0382 | Not recorded | 69857574c97a0b3e3b5d5b84 | | FireRedTTS | 0.0382 | Not recorded | 6a0beeb9dbc836b87d82c596 | | CosyVoice | 0.0429 | Not recorded | 6a0beeb9dbc836b87d82c596 | | HoliTok-Unite | 0.072 | Not recorded | 6a191ddcdbd7150506f82b74 | | HoliTok-Base | 0.2785 | Not recorded | 6a191ddcdbd7150506f82b74 | | MingTok-Audio | 0.5106 | Not recorded | 6a191ddcdbd7150506f82b74 | | Semantic-VAE | 1.0232 | Not recorded | 6a191ddcdbd7150506f82b74 | ### ARCHI-TTS WER: 0.0147. Result ID: 699a108d5698665ec49437d6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### DiTAR WER: 0.0169. Result ID: 699a108d5698665ec49437d5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### SemaVoice WER: 0.0171. Result ID: 6a0c28e4fbc373592c027046. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"C-AR","Params":"1.5B","#Hours":"150K"}. This record does not assert independent reproduction. ### Seed-TTS\_DiT WER: 0.01733. Result ID: 699a108d5698665ec49437d2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### F5-TTS WER: 0.0183. Result ID: 699a108d5698665ec49437d4. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### VoxCPM WER: 0.0185. Result ID: 6a0c28e4fbc373592c027043. [Reporting paper]() Recorded evaluation configuration: {"Type":"C-AR","Params":"0.5B","#Hours":"1.8M"}. This record does not assert independent reproduction. ### SemaVoice-Emilia WER: 0.0191. Result ID: 6a0c28e4fbc373592c027045. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"C-AR","Params":"1.5B","#Hours":"100K"}. This record does not assert independent reproduction. ### OpenAudio-s1 WER: 0.0194. Result ID: 6a0c28e4fbc373592c02703c. [Reporting paper]() Recorded evaluation configuration: {"Type":"D-AR","Params":"0.5B","#Hours":"2.0M"}. This record does not assert independent reproduction. ### FireRedTTS-2 WER: 0.0195. Result ID: 6a0c28e4fbc373592c02703b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"D-AR","Params":"-","#Hours":"1.4M"}. This record does not assert independent reproduction. ### SparkTTS WER: 0.0198. Result ID: 6a0c28e4fbc373592c02703a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"D-AR","Params":"0.5B","#Hours":"100K"}. This record does not assert independent reproduction. ### F5-TTS WER: 0.02. Result ID: 6a0c28e4fbc373592c027038. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Type":"C-NAR","Params":"0.3B","#Hours":"100K"}. This record does not assert independent reproduction. ### Ground Truth WER: 0.0206. Result ID: 699a108d5698665ec49437cc. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### VAE Resynthesized WER: 0.0206. Result ID: 699a108d5698665ec49437cd. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Vocos Resynthesized WER: 0.0209. Result ID: 699a108d5698665ec49437ce. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Ground Truth WER: 0.0214. Result ID: 6a0c28e4fbc373592c027036. [Reporting paper]() Recorded evaluation configuration: {"Type":"-","Params":"-","#Hours":"-"}. This record does not assert independent reproduction. ### E2 TTS WER: 0.0219. Result ID: 699a108d5698665ec49437d3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### IndexTTS 2 WER: 0.0223. Result ID: 6a0c28e4fbc373592c027041. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"D-AR+C-NAR","Params":"1.5B","#Hours":"55K"}. This record does not assert independent reproduction. ### VoxCPM-Emilia WER: 0.0234. Result ID: 6a0c28e4fbc373592c027042. [Reporting paper]() Recorded evaluation configuration: {"Type":"C-AR","Params":"0.5B","#Hours":"100K"}. This record does not assert independent reproduction. ### HiggsAudio-v2 WER: 0.0244. Result ID: 6a0c28e4fbc373592c02703d. [Reporting paper]() Recorded evaluation configuration: {"Type":"D-AR","Params":"3.0B","#Hours":"10M"}. This record does not assert independent reproduction. ### CosyVoice 2 WER: 0.0257. Result ID: 699a108d5698665ec49437cf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### CosyVoice2 WER: 0.0257. Result ID: 6a0c28e4fbc373592c02703f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"D-AR+C-NAR","Params":"0.5B","#Hours":"170K"}. This record does not assert independent reproduction. ### MaskGCT WER: 0.0262. Result ID: 6a0c28e4fbc373592c027039. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"D-NAR","Params":"1.0B","#Hours":"100K"}. This record does not assert independent reproduction. ### MaskGCT WER: 0.02623. Result ID: 699a108d5698665ec49437d1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Qwen2.5-Omni WER: 0.0272. Result ID: 6a0c28e4fbc373592c027037. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"MLLM","Params":"7.0B","#Hours":"-"}. This record does not assert independent reproduction. ### VibeVoice WER: 0.0304. Result ID: 6a0c28e4fbc373592c027044. [Reporting paper]() Recorded evaluation configuration: {"Type":"C-AR","Params":"1.5B","#Hours":"-"}. This record does not assert independent reproduction. ### FireRedTTS WER: 0.0382. Result ID: 699a108d5698665ec49437d0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### FireRedTTS WER: 0.0382. Result ID: 6a0c28e4fbc373592c027040. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"D-AR+C-NAR","Params":"0.5B","#Hours":"248K"}. This record does not assert independent reproduction. ### CosyVoice WER: 0.0429. Result ID: 6a0c28e4fbc373592c02703e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Type":"D-AR+C-NAR","Params":"0.3B","#Hours":"170K"}. This record does not assert independent reproduction. ### HoliTok-Unite WER: 0.072. Result ID: 6a196b5dfbc373592c146d22. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modeling":"Unified spoken language modeling"}. This record does not assert independent reproduction. ### HoliTok-Base WER: 0.2785. Result ID: 6a196b5dfbc373592c146d21. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modeling":"Unified spoken language modeling"}. This record does not assert independent reproduction. ### MingTok-Audio WER: 0.5106. Result ID: 6a196b5dfbc373592c146d20. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modeling":"Unified spoken language modeling"}. This record does not assert independent reproduction. ### Semantic-VAE WER: 1.0232. Result ID: 6a196b5dfbc373592c146d1f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modeling":"Unified spoken language modeling"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.