AlibabaQwen ASR
Qwen Audio 3.1 ASR Flash
Qwen Audio 3.1 ASR Flash has four interfaces: short-message WebSocket, continuous WebSocket, HTTP recognition and offline file transcription.
Practical use
Practical guidance. Choose the interface before designing the transcript workflow. Keep live partial text provisional until the final result arrives. For recordings, review speaker assignments and chunk boundaries separately from ordinary word accuracy.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API |
| Model identifiers | qwen-audio-3.1-asr-flash-messageqwen-audio-3.1-asr-flash-streamingqwen-audio-3.1-asr-flash-filetransqwen-audio-3.1-asr-flash |
| Short messages | Optional text polishing can revise the final transcript.Speech recognition model catalogue |
| Files | Filetrans supports speaker separation and recordings up to 12 hours or 2 GB.Speech recognition model catalogue |
| Speaker separation | Available in HTTP Flash and Filetrans, not the message or streaming variants.Speech recognition model catalogue |
Related models
- Qwen3-ASR-0.6B
Alibaba · Qwen ASR
Compact Qwen3-Omni-based speech recognizer with streaming and offline inference.
Launch date not confirmedDownloadable - Qwen3-ASR-1.7B
Alibaba · Qwen ASR
Speech recognition model for streaming and recorded audio in 30 languages and 22 Chinese dialects.
Launch date not confirmedDownloadable - Qwen3-ForcedAligner-0.6B
Alibaba · Qwen ASR
Speech alignment model that assigns timestamps to supplied text in 11 languages.
Launch date not confirmedDownloadable