AlibabaQwen ASR
Qwen3-ASR-1.7B
Qwen3-ASR-1.7B shares the streaming and offline inference design of the smaller model, with 30 languages and 22 Chinese dialects.
Practical use
Practical guidance. Compare long recordings as well as short clips. Track omissions and repeated text across chunk boundaries, then decide whether the accuracy difference justifies the deployment resources.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | Qwen/Qwen3-ASR-1.7B |
| Ancestry | Built from Qwen3-Omni audio capabilities.Qwen3-ASR-1.7B model card |
| Licence | Apache 2.0.Qwen3-ASR-1.7B model card |
Related models
- Qwen3-ASR-0.6B
Alibaba · Qwen ASR
Compact Qwen3-Omni-based speech recognizer with streaming and offline inference.
Launch date not confirmedDownloadable - Qwen Audio 3.1 ASR Flash
Alibaba · Qwen ASR
Speech recognition model for short messages, continuous streams and recorded files.
Launch date not confirmedAPI - Qwen3-ForcedAligner-0.6B
Alibaba · Qwen ASR
Speech alignment model that assigns timestamps to supplied text in 11 languages.
Launch date not confirmedDownloadable