A scalable multilingual zero-shot text-to-speech synthesizer based on large language models, designed for speech generation in the wild.
State-of-the-art TTS with powerful capabilities
Covers 9 languages (Chinese, English, Japanese, Korean, German, Spanish, French, Italian, Russian) and 18+ Chinese dialects including Cantonese, Minnan, Sichuan, Shanghai, and more.
Clone any voice with just a short audio prompt. Supports both multi-lingual and cross-lingual zero-shot synthesis with high speaker similarity.
Supports both text-in streaming and audio-out streaming, achieving latency as low as 150ms while maintaining high-quality audio output.
Supports various instructions for controlling languages, dialects, emotions, speaking speed, volume, and more for fine-grained speech generation.
Built-in reading of numbers, special symbols, and various text formats without requiring a traditional frontend module.
Supports Chinese Pinyin and English CMU phoneme inpainting for precise pronunciation control, suitable for production use.
Listen to samples generated by each version of CosyVoice
Latest version with SOTA performance in content consistency, speaker similarity, and prosody naturalness. Supports 9 languages and 18+ Chinese dialects.
LatestScalable streaming speech synthesis with large language models. Features 25Hz codec and improved streaming inference.
StreamingThe original release featuring zero-shot, cross-lingual, instruct, and SFT inference modes with multilingual support.
FoundationGet CosyVoice running in minutes
Benchmark results on CV3-Eval test sets
| Model | Open Source | Size | test-zh CER%↓ | test-zh SS%↑ | test-en WER%↓ | test-en SS%↑ | test-hard CER%↓ | test-hard SS%↑ |
|---|---|---|---|---|---|---|---|---|
| Human | - | - | 1.26 | 75.5 | 2.14 | 73.4 | - | - |
| Seed-TTS | ❌ | - | 1.12 | 79.6 | 2.25 | 76.2 | 7.59 | 77.6 |
| F5-TTS | ✅ | 0.3B | 1.52 | 74.1 | 2.00 | 64.7 | 8.67 | 71.3 |
| CosyVoice2 | ✅ | 0.5B | 1.45 | 75.7 | 2.57 | 65.9 | 6.83 | 72.4 |
| Index-TTS2 | ✅ | 1.5B | 1.03 | 76.5 | 2.23 | 70.6 | 7.12 | 75.5 |
| VoxCPM | ✅ | 0.5B | 0.93 | 77.2 | 1.85 | 72.9 | 8.87 | 73.0 |
| Fun-CosyVoice3-0.5B | ✅ | 0.5B | 1.21 | 78.0 | 2.24 | 71.8 | 6.71 | 75.8 |
| Fun-CosyVoice3-0.5B RL | ✅ | 0.5B | 0.81 | 77.4 | 1.68 | 69.5 | 5.44 | 75.0 |
Download from ModelScope or HuggingFace
Latest and recommended model with best overall performance. 0.5B parameters.
Streaming TTS with 25Hz codec. Supports real-time speech synthesis.
CosyVoice is part of a complete speech AI toolkit
Industrial speech recognition with 50+ languages, speaker diarization, and streaming support.
End-to-end LLM-based ASR supporting 31 languages with hotwords and vLLM streaming.
Ultra-fast ASR with emotion detection and audio event recognition.
AI-powered video clipping driven by speech recognition.
Read the research behind CosyVoice