Voicv

CosyVoice is Qwen Cloud’s voice cloning and speech synthesis model in the CosyVoice series, designed for professional text-to-speech scenarios with improved sound quality, naturalness, expressiveness, and cloning fidelity. With a short reference recording, it can create a highly similar custom voice without model training; Qwen recommends 10–20 seconds of clear speech, while at least five seconds of continuous speech is required. The model supports real-time, streaming text-to-speech synthesis, allowing applications to accept text and return audio with low first-packet latency. It supports Chinese, English, French, German, Japanese, Korean, and Russian for cloned voices, with language hints available to improve identification during enrollment. Source recordings can use WAV, MP3, or M4A formats and should contain clean speech without background music, noise, or additional speakers.

Qwen3-TTS is an open source series of advanced text-to-speech models developed by the Qwen team at Alibaba Cloud under the Apache-2.0 license, offering stable, expressive, and real-time speech generation with features such as voice cloning, voice design, and fine-grained control of prosody and acoustic attributes. The models support 10 major languages, including Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish, and Italian, and multiple dialectal voice profiles with adaptive control over tone, speaking rate, and emotional expression based on text semantics and instructions. Qwen3-TTS uses efficient tokenization and a dual-track architecture that enables ultra-low-latency streaming synthesis (first audio packet in ~97 ms), making it suitable for interactive and real-time use cases, and includes a range of models with different capabilities (e.g., rapid 3-second voice cloning, custom voice timbres, and instruction-based voice design).