Workflow Generation on DRFLOW (original)
94.92FactualityDRFA
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| DRFABackbone=DeepSeek-v3.22026.06 | 94.92 | 91.74 | 35.85 | 56.32 | 68.4 | 48.45 | 55.1 | |
| DRBABackbone=DeepSeek-v3.22026.06 | 93.57 | 89.43 | 15.02 | 33.51 | 60.48 | 38.89 | 45.37 | |
| DRBABackbone=Gemini-3.1-pro2026.06 | 90.24 | 83.12 | 14.9 | 30.38 | 55.94 | 37.07 | 40.99 | |
| DRBABackbone=GPT-5.22026.06 | 90.01 | 89.63 | 24.85 | 44.99 | 60.43 | 52.11 | 54.17 | |
| DRBABackbone=Average2026.06 | 89.8 | 87.66 | 20.19 | 37.69 | 60.51 | 43.21 | 47.91 | |
| DRFABackbone=Average2026.06 | 87.84 | 90.82 | 38.21 | 56.54 | 69.3 | 52.31 | 57.93 | |
| DRFABackbone=Gemini-3.1-pro2026.06 | 87.16 | 91.2 | 26.98 | 48.08 | 67.31 | 42.55 | 52.05 | |
| DRFABackbone=GPT-5.22026.06 | 85.99 | 92.22 | 42.83 | 59.26 | 69.43 | 62.34 | 63.9 | |
| DRBABackbone=Claude-Opus-4.52026.06 | 85.36 | 88.47 | 25.99 | 41.89 | 65.19 | 44.76 | 51.09 | |
| DRFABackbone=Claude-Opus-4.52026.06 | 83.31 | 88.13 | 47.18 | 62.5 | 72.07 | 55.87 | 60.7 | |
| BasicBackbone=GPT-5.22026.06 | 59.46 | 47.5 | 2.86 | 12.86 | 24.52 | 9.79 | 12.16 | |
| BasicBackbone=Claude-Opus-4.52026.06 | 50.1 | 70.07 | 1.64 | 10.29 | 27.93 | 10.62 | 13.54 | |
| BasicBackbone=Average2026.06 | 47.99 | 52.72 | 1.45 | 9.77 | 21.25 | 8.04 | 9.84 | |
| BasicBackbone=DeepSeek-v3.22026.06 | 43.8 | 55.33 | 1.17 | 9.45 | 19.04 | 6.97 | 8.61 | |
| BasicBackbone=Gemini-3.1-pro2026.06 | 38.59 | 38 | 0.11 | 6.49 | 13.51 | 4.77 | 5.05 |