Alibaba cuts AI voice service costs by up to 95% for underserved regions

Alibaba’s Qwen-Audio-3.1 release slashes voice AI prices for low-cost communication in underserved areas

Alibaba launched Qwen-Audio-3.1 on September 23, 2026, with five new models: an ASR model improving multilingual and dialect recognition, an ASR-Next model for multi-speaker identification with emotion detection, a TTS model enabling multilingual voice synthesis, a TTS-Next model using diffusion approaches for sound generation, and a real-time interaction model handling interruptions. The release cuts TTS service prices by 70%, real-time services by 85%, and ASR services by up to 95%. This makes voice-based AI tools significantly cheaper for low-resource settings, potentially expanding access to information and communication in underserved regions. Price reductions apply to specific service tiers and are subject to Qwen Cloud’s terms.

Source: The Decoder