Web Agent Tasks on WorkArena-L1 (test)
55.56Success Rate (SR)Vanilla-IB
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Vanilla-IBBackbone=Qwen 3.6-27B2026.06 | 55.56 | 109.4 | |
| ReasoningBankBackbone=Qwen 3.6-27B2026.06 | 55.56 | 120.3 | |
| AWMBackbone=Qwen 3.6-27B2026.06 | 53.53 | 102.8 | |
| ASIBackbone=Qwen 3.6-27B2026.06 | 48.49 | 113.8 |