Release dates
— QwenCloud released the distinct `qwen-audio-3.1-realtime-flash` model for duplex voice with text/audio input and output, function calling, and web search. ↗
Qwen-Audio 3.1 Realtime Flash
The full-duplex realtime speech Flash model in Qwen-Audio 3.1. It accepts text or audio and returns text and speech, with function calling, web search, and voice cloning.
QwenCloud released API model `qwen-audio-3.1-realtime-flash` on 2026-10-27. Model Studio documentation lists Singapore and Beijing regions and WebSocket, WebRTC, and AOQ access.
MODEL NUMBER153
- First public date
- 2026-10-27
- Organization
- Alibaba Cloud
- Family
- Qwen
01
Specifications
Only published specifications are shown.
02
Overview
Why it mattered
This separately named API model extends the Qwen-Audio 3.1 Realtime line; the originator recommends the Flash version for cost-sensitive realtime voice conversations. It is kept as a specialized timeline entry, not merged with Realtime Plus, and no weight or architecture difference is inferred.
03
Architecture & capabilities
Model capabilities
Native audioReal-time audioTool useVoice cloning
Connected tools
Web search
API features
Function calling
04
Release history
QwenCloud released the distinct `qwen-audio-3.1-realtime-flash` model for duplex voice with text/audio input and output, function calling, and web search.
View source05
Related models
Select a card to view the related model.
06
Sources
01
QwenCloud model releases: qwen-audio-3.1-realtime-flashQwenCloud / Alibaba Cloud · Checked 2026-10-28
Official · Verified02Qwen-Audio 3.1 Realtime Flash model guideAlibaba Cloud Model Studio · Checked 2026-10-28
Official · Verified03QwenCloud speech-to-speech model selectionQwenCloud / Alibaba Cloud · Checked 2026-10-28
Official · VerifiedSuggest a correction
Copy this template, add the field, proposed value, and primary source, and share it with the archive maintainer.