Operational Tasks on Internal offline website benchmark
71.8Task Completion Rate (TCR)WebFactory-3B
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| WebFactory-3B2026.03 | 71.8 | 0.82 | 87.6 | |
| GUI-R1-3B2026.03 | 68.2 | 0.78 | 85.3 | |
| GPT-4o2026.03 | 26.7 | 0.41 | 48.6 | |
| QwenVL2.5-3B2026.03 | 18.3 | 0.32 | 41.2 |