Inference · text to speech

English text to speech.

Every open model that has done text to speech in English on the network, fastest first. Speed is how many times faster than real time the machine produced it; below 1× it is slower than real time.

3models
ModelFamilySpeedMachines
kokoro baseKokoro1.7× real time6
F5-TTS v1 Base (zero-shot voice, en/zh)F5-TTS0.3× real time1
MiniCPM-o 4.5 BASEMiniCPM-o 4.5measured2