Personalized workflow prediction on 15 tasks ablation
96.66FactualityDRBA
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| DRBAModel=GPT-5.52026.06 | 96.66 | 88.5 | 35.46 | 46.22 | 66.29 | 63.75 | 64.78 | |
| DRBAModel=Kimi-K2.62026.06 | 92.26 | 87.55 | 34.84 | 45.25 | 62.3 | 65.48 | 62.16 | |
| DRBAModel=GLM-5.12026.06 | 88.55 | 86.96 | 33.5 | 40.16 | 62.02 | 64.98 | 62.3 | |
| DRFAModel=GLM-5.12026.06 | 86.19 | 86.18 | 46.91 | 59.59 | 70.73 | 65.36 | 66.57 | |
| DRFAModel=GPT-5.52026.06 | 85.92 | 90.37 | 48.94 | 65.55 | 71.34 | 67.22 | 68.62 | |
| DRBAModel=Claude-Opus-4.72026.06 | 83.55 | 89.44 | 38.59 | 50.98 | 69.06 | 64.76 | 65.19 | |
| DRFAModel=Kimi-K2.62026.06 | 83.51 | 86.41 | 49.71 | 59.38 | 70.87 | 68.74 | 68.54 | |
| DRFAModel=Claude-Opus-4.72026.06 | 82.63 | 91.24 | 51.83 | 65.67 | 73.11 | 68.6 | 69.14 | |
| DRFAModel=Qwen3.6-plus2026.06 | 82.3 | 88.51 | 43.39 | 52.83 | 68.14 | 68.54 | 67.42 | |
| DRBAModel=Qwen3.6-plus2026.06 | 77.65 | 88.45 | 27.4 | 38.42 | 59.23 | 65.75 | 61.4 |