User Satisfaction Prediction on PersTurnBench (test)
4.737Micro Satisfaction Scorekimi-k2.6
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| kimi-k2.62026.05 | 4.737 | 4.812 | 4.731 | 4.785 | 98.7 | 1.3 | |
| glm-5.12026.05 | 4.707 | 4.8 | 4.697 | 4.76 | 97.3 | 2.7 | |
| deepseek-v4-pro2026.05 | 4.667 | 4.791 | 4.655 | 4.732 | 97.3 | 2.7 | |
| gpt-5.52026.05 | 4.573 | 4.689 | 4.564 | 4.624 | 94.3 | 5.7 | |
| claude-opus-4-72026.05 | 4.54 | 4.636 | 4.524 | 4.597 | 95.3 | 4.7 | |
| gemini-3.1-pro2026.05 | 4.527 | 4.628 | 4.532 | 4.583 | 94 | 6 | |
| minimax-m2.72026.05 | 4.263 | 4.426 | 4.252 | 4.344 | 86.3 | 13.7 |