Alibaba has released Qwen-Audio-3.1-Realtime, a full-duplex voice model designed to enhance interactive communication. This innovative model is trained to reason, call tools, and determine the optimal moments to speak, significantly improving user interaction.
In testing, the Qwen-Audio-3.1-Realtime model demonstrated a task success rate of 82.0%, an increase from 78.4% achieved by its predecessors. Furthermore, the model effectively minimizes interruptions from background speech, reducing responses to such distractions from 73% to just 13%. This advancement is now available as an API on QwenCloud, marking a significant step forward in voice technology.
Compiled automatically by the Tech AI Newsdesk from public AI-news sources and summarised in our own words.
Frequently asked questions
What is Qwen-Audio-3.1-Realtime?
Qwen-Audio-3.1-Realtime is a full-duplex voice model released by Alibaba, designed to enhance interactive communication.
How does Qwen-Audio-3.1-Realtime improve user interaction?
The model improves user interaction by reasoning, calling tools, and determining optimal moments to speak, achieving a task success rate of 82.0%.
