Audio Instruction Following on VoiceBench
4.78AlpacaEval ScoreGPT-4o-Audio
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| GPT-4o-Audio2025.07 | 4.78 | 4.49 | 4.58 | 75.5 | 80.25 | 89.23 | 84.1 | 76.02 | 98.65 | 86.75 | |
| ASR+LLM2025.07 | 4.53 | 4.04 | 4.16 | 70.43 | 62.43 | 72.53 | 69.7 | 69.53 | 98.08 | 77.48 | |
| DeSTA2.5-Audio2025.07 | 4.41 | 3.99 | 3.8 | 60.04 | 60.87 | 74.06 | 67 | 66.42 | 98.27 | 74.52 | |
| DeSTA22025.07 | 4.36 | 3.33 | 3.42 | 56.06 | 66.43 | 61.98 | 66.2 | 52.63 | 98.27 | 69.31 | |
| VITA-1.52025.07 | 4.21 | 3.66 | 3.48 | 38.88 | 52.15 | 71.65 | 55.3 | 38.14 | 97.69 | 64.53 | |
| GLM-4-Voice2025.07 | 3.97 | 3.42 | 3.18 | 36.98 | 39.75 | 53.41 | 52.8 | 25.92 | 88.08 | 56.48 | |
| Phi-42025.07 | 3.81 | 3.82 | 3.56 | 39.78 | 42.19 | 65.93 | 61.8 | 45.35 | 100 | 64.32 | |
| Qwen2-Audio-Instruct2025.07 | 3.74 | 3.43 | 3.01 | 35.71 | 35.72 | 49.45 | 54.7 | 26.33 | 96.73 | 55.8 | |
| DiVA2025.07 | 3.67 | 3.54 | 3.74 | 57.05 | 25.76 | 25.49 | 51.8 | 39.15 | 98.27 | 57.39 | |
| Moshi2025.07 | 2.01 | 1.6 | 1.3 | 15.64 | 24.04 | 25.93 | 47.4 | 10.12 | 44.23 | 29.51 |