Alibaba Cloud Model Studio:Reconnaissance vocale non temps réel (Qwen-Audio-3.0-ASR-Flash-Filetrans/Fun-ASR)