Alibaba Tongyi Audio (Qwen-Audio)
Alibaba's Tongyi multimodal model exposed for transcription + audio understanding.
Alibaba's Tongyi multimodal model exposed for transcription + audio understanding.
Best for mandarin-heavy multimodal workflows that want one foundation model for audio + reasoning. Pricing: tiered RMB-per-token / per-second.
What it is
Tongyi is Alibaba's flagship LLM family; Qwen-Audio is the audio-understanding variant exposed via DashScope. In addition to plain transcription, it provides audio-grounded question answering, captioning, and content classification. Strongest in Mandarin. China-first; international tenants should verify region availability. Best fit: mandarin-heavy multimodal workflows that want one foundation model for audio + reasoning. Caveats: china-first; international availability via alibaba cloud international. Pricing as listed: tiered RMB-per-token / per-second. Directory tags: commercial-api, regional-asia, foundation-model. Last vendor-page check: 2026-05-12. The product slots into the broader voice-AI tooling landscape and is best evaluated head-to-head with adjacent vendors in its subcategory; verify current language coverage, region availability, and compliance terms on the vendor's public docs at time of build.
Watch out for: China-first; international availability via Alibaba Cloud International.
Install / use
Alibaba Cloud DashScope: Qwen-Audio familyFeatures
| Speaker diarization | No |
| Word-level timestamps | No |
| Streaming / real-time | No |
| Languages supported | None |
| HIPAA eligible | No |
Alibaba Tongyi Audio (Qwen-Audio) vs Whipscribe
| Feature | Alibaba Tongyi Audio (Qwen-Audio) | Whipscribe |
|---|---|---|
| Category | Transcription APIs | Transcription APIs |
| Pricing | tiered RMB-per-token / per-second | free beta |
| Speaker diarization | — | Yes |
| Word timestamps | — | Yes |
| Streaming | — | No |
| Languages | — | 99 |
| Platforms | API | Web, API, MCP |
Alternatives to Alibaba Tongyi Audio (Qwen-Audio)
Whipscribe is a managed faster-whisper + whisperX service. If you want transcripts without running infrastructure, paste a URL or drop a file in the form below — you'll have a transcript in seconds.