Empathy Response Generation on VStyle (test)
4.98Anger Score (en)GPT-4o-Mini
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| GPT-4o-Mini2026.02 | 4.98 | 4.3 | 5 | 4.52 | 4.87 | 4.73 | 5 | 4.69 | |
| GPT-4o2026.02 | 4.95 | 4.75 | 4.9 | 4.83 | 5 | 4.67 | 4.54 | 4.8 | |
| Doubao2026.02 | 4.89 | 4.59 | 5 | 4.72 | 4.81 | 4.8 | 4.94 | 4.83 | |
| Covo-Audio-Chat2026.02 | 4.86 | 4.89 | 4.83 | 4.93 | 4.97 | 5 | 4.8 | 4.89 | |
| Step-Audio2026.02 | 3.95 | 3.59 | 4.37 | 4.52 | 3.87 | 3.2 | 4.29 | 4.26 | |
| Fun-Audio-Chat-8B2026.02 | 3.64 | 3.73 | 4.1 | 3.93 | 2.9 | 4.03 | 3.69 | 3.77 | |
| Kimi-Audio2026.02 | 3.59 | 3.86 | 3.97 | 3.86 | 3.65 | 3.8 | 3.46 | 4.57 | |
| Qwen2.5-Omni2026.02 | 2.95 | 4.64 | 2.73 | 4.28 | 3.55 | 4.77 | 3.43 | 4.91 | |
| Baichuan-Audio2026.02 | 2.41 | 2.11 | 3.43 | 2.55 | 2.74 | 2.2 | 3.91 | 3.51 |