Alibaba launches voice recognition model 'Qwen-Audio-3.0-ASR-Flash'... "Improved accuracy for technical terms"

Alibaba has launched its new speech recognition models, Qwen-Audio-3.0-ASR-Flash and Streaming.

The models focus on long-form audio, specialized terminology, and custom vocabulary recognition.

They also support document-style outputs, multilingual use, and low-latency real-time transcription.

Alibaba says the models are available as APIs on its Cloud Bailian platform.

Potential use cases include healthcare, industrial settings, IT, meeting notes, and customer support.