Alibaba's Qwen releases upgraded audio ASR model with enhanced domain recognition

Alibaba's Qwen team has released Qwen-Audio-3.0-ASR-Flash, an upgraded automatic speech recognition model. This new version offers improved context awareness and domain-term recognition, with enhancements for custom hotwords and speech polishing into structured transcripts. Internal testing showed high recall rates, with 95.36% for medical terms and 93.24% for industrial terms. AI

IMPACT Enhances specialized audio transcription capabilities, potentially improving AI applications in fields like medicine and industry.

RANK_REASON Frontier-lab model release with system card [lever_c_demoted from frontier_release: ic=1 ai=1.0]

Read on X — Qwen (Alibaba) →

• Alibaba Group
• Qwen
• Qwen-Audio-3.0-ASR-Flash
• Qwen-Audio-3.0-ASR-Flash-Filetrans
• Qwen-Audio-3.0-ASR-Flash-Streaming

• model release
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →