Vendor / paper demo
Speaker 1 (zero-shot) · CosyVoice 2 clip from the IndexTTS comparison table, Speaker 1.
Old will is a fine fellow but poor and helpless since missus rogers had her accident.
FunAudioLLM / Alibaba · 2024–26 · CosyVoice 2/3
CosyVoice 2
A 0.5B streaming zero-shot model from FunAudioLLM. CosyVoice 3 continues the line. The clip is Speaker 1 from the IndexTTS comparison page (same sentence as XTTS, IndexTTS, F5, Fish Speech).