Alibaba Qwen-Audio-3.0 Speech Models

Overview

Alibaba announced the full launch of the Qwen-Audio-3.0 series on the Qwen AI platform in mid-August 2026, describing it as a suite of speech models that achieved top rankings across major international evaluations.[15] The series targets tasks such as speech recognition, translation and synthesis and is integrated with Tongyi Qianwen, enabling developers to build end-to-end voice interfaces atop the broader Qwen ecosystem.[15] The focus is on high accuracy, robustness in noisy environments and coverage of multiple languages including Chinese and English.[15]