Web Agent Task Completion on GAIA
70.9Accuracy
SlimSearcher (SFT+RL)
Updated 1mo ago
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
2026.06 | 70.9 | 10.61 | 4,915 | |
2026.06 | 70.2 | 16.65 | 5,045 | |
2026.06 | 69.9 | 20.13 | 7,237 | |
2026.06 | 68.3 | - | - | |
2026.06 | 68.2 | 20.56 | 7,378 | |
2026.06 | 67.4 | - | - | |
2026.06 | 66.5 | 24.46 | 7,299 | |
2026.06 | 66.3 | 18.7 | 6,752 | |
2026.06 | 66 | 21.79 | 6,412 | |
2026.06 | 46.9 | 6.45 | 1,281 | |
2026.06 | 45.6 | 7.14 | 1,128 | |
2026.06 | 37.3 | 3.2 | 994 | |
2026.06 | 37.2 | 12.88 | 3,560 | |
2026.06 | 33.9 | 11.63 | 986 |