Alibaba launches Qwen Audio 3.1 with new models and slashes AI audio prices by up to 95 percent
decoder ·
Alibaba's Qwen team has released Qwen-Audio-3.1, a suite of five new AI models designed for speech recognition, text-to-speech synthesis, and real-time interaction. The updated Automatic Speech Recognition (ASR) capabilities now include improved multilingual support, dialect handling, and the ability to filter out filler words while identifying multiple speakers and emotions. For text-to-speech, the new TTS-Next model utilizes a diffusion approach to generate voice, sound effects, and background audio in a single pass, allowing users to control tone and style via simple text prompts. Additionally, the real-time interaction model enables simultaneous speaking and listening, adapting its response speed and empathy based on detected user mood. Alongside these technical advancements, Alibaba is significantly reducing costs for developers, with price cuts reaching up to 95 percent for ASR services. Text-to-speech pricing has dropped by approximately 70 percent, while real-time interaction costs have decreased by roughly 85 percent. These updates are available through the Qwen Cloud platform, offering enhanced functionality and affordability for audio AI applications.
decoder