VoxCPM Evaluation Results
December 26, 2025 · View on GitHub
Model: VoxCPM-0.5B Evaluation Date: 2025/12/08 Paper/Repo: openbmb/VoxCPM-0.5B
Metrics Legend:
- WER⬇️: Word Error Rate (lower is better)
- CER⬇️: Character Error Rate (lower is better)
- SIM⬆️: Speaker Similarity (higher is better)
- P808_MOS⬆️: DNSMOS P.808 Mean Opinion Score (higher is better)
Seed-TTS-Eval Benchmark
| task | dataset | WER/CER⬇️ | SIM⬆️ | eval_cli | note |
|---|---|---|---|---|---|
| tts | seed_tts_eval_en | 1.81(1.85) | 73.23(72.9) | [1] | |
| tts | seed_tts_eval_zh | 1.04(0.93) | 77.10(77.2) | [2] |
CV3 Benchmark (Zero-Shot)
| task | dataset | WER/CER⬇️ | SIM⬆️ | P808_MOS⬆️ | eval_cli | note |
|---|---|---|---|---|---|---|
| tts | cv3_zero_shot_en | 6.64(4.04) | 67.93 | 3.78 | [3] | |
| tts | cv3_zero_shot_zh | 3.40(3.40) | 71.84 | 3.85 | [4] | |
| tts | cv3_zero_shot_hard_en | 7.45(7.89) | 64.99(64.3) | 3.93(3.74) | [5] | |
| tts | cv3_zero_shot_hard_zh | 9.95(12.9) | 66.78(66.1) | 3.74(3.59) | [6] |
Evaluation Commands
[1] python audio_evals/main.py --dataset seed_tts_eval_en --model voxcpm-vc
[2] python audio_evals/main.py --dataset seed_tts_eval_zh --model voxcpm-vc
[3] python audio_evals/main.py --dataset cv3_zero_shot_en --model voxcpm-vc
[4] python audio_evals/main.py --dataset cv3_zero_shot_zh --model voxcpm-vc
[5] python audio_evals/main.py --dataset cv3_zero_shot_hard_en --model voxcpm-vc
[6] python audio_evals/main.py --dataset cv3_zero_shot_hard_zh --model voxcpm-vc