The Qwen3-TTS series introduces open-source models for speech generation, voice design, and cloning, available in 1.7B and 0.6B sizes. These models support 10 languages and offer features like rapid voice cloning and style control. They excel in multilingual capabilities and efficient speech signal processing.