Adjustable Sample Rate
48000 Hz44100 Hz24000 Hz16000 Hz12000 Hz8000 Hz
Download After Generation
Emotion Tags
normalslowcryingsleepysighchuckle
Guided Emotion and Intonation
Input Streaming for Lower Latency
LLM-based Customizability
Maximum Text Length Per Query
2000000
Model Tokenizer Type
Non-streaming (CNN-based) tokenizer
Open Source Release Planned
Orpheus Speech Models
Medium (3B)Small (1B)Tiny (400M)Nano (150M)
Pretrained and Finetuned Models
Pretrained modelsFinetuned models
Sample Finetuning Scripts
Sliding Window Detokenizer
Streaming Inference Speed
Faster than playback on A100 40GB for 3B model
Total Supported Languages
146
Training Data Volume
100k+ hours of speech, billions of text tokens
Voice Selection
AmeliaAndrewAvaBlueBrian WhiteDanielle plusDavisElijahEmma CoxEvelynGregory plusHollyHunterIvy plusJane SmithJasonJoanna plusJoey plusJohnJustin plusKaiKendra plusKimberly plusKolbyLuna BellMatthew plusNancyOliverRogerRuth plusSalli plusScottSimon USSkylerSteffanTonyJenniferHannaKelsyArnoldEllisJackFennyBartDenAngelGloryHelenIronJerryAveryGuyAmberAnnyAshleyBrandonChristopherCoraElizabethEricJacobJennyMichelleMonicaSaraKevin plusSalliJoannaIvyKendraKimberlyMatthewJustinJoeyAda USAdam USAlessio USAlloy USAmanda USAndrew USArabella USAva USBrandon USBrian USChristopher USCora USDavis USDerek USDustin USEcho USEmma USEvelyn USFable USFlorian USGiuseppe USHyunsu USIsabella USIsidora USJenny USLewis USLola USLucien USMacerio USMarcello USMasaru USNancy USNova USOllie USOnyx USPhoebe USRemy USRyan USSamuel USSeraphina USSerena USShimmer USSteffan USThalita USTristan USVivienne USXiaochen USXiaoxiao USXiaoyu USXimena USYunfan USYunxiao USYunyi US