Resource Hijacking Evaluation on GAIA (test)
114,725Research And DataNormal
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| NormalAgent=Cloud-Hosted (OpenManus)2025.12 | 114,725 | 148,609 | 84,192 | 105,747 | 117,802 | |
| LeechHijackAgent=Cloud-Hosted (OpenManus)2025.12 | 106,902 | 157,260 | 87,856 | 109,390 | 120,417 | |
| LeechHijackAgent=Localized (OpenManus)2025.12 | 105,170 | 170,331 | 102,229 | 81,470 | 123,627 | |
| LeechHijackAgent=Hybrid (OWL)2025.12 | 102,703 | 125,296 | 100,420 | 102,493 | 108,755 | |
| LeechHijackAgent=Hybrid (Pydantic-AI)2025.12 | 97,867 | 114,099 | 104,216 | 115,342 | 91,357 | |
| NormalAgent=Hybrid (Pydantic-AI)2025.12 | 89,540 | 135,758 | 91,846 | 103,716 | 98,333 | |
| NormalAgent=Localized (OpenManus)2025.12 | 85,286 | 161,374 | 77,610 | 65,832 | 113,906 | |
| NormalAgent=Hybrid (OWL)2025.12 | 79,206 | 109,707 | 72,810 | 75,480 | 95,118 | |
| Extra TaskAgent=Hybrid (OWL)2025.12 | 16,726 | 7,212 | 10,490 | 14,634 | 22,174 | |
| Extra TaskAgent=Localized (OpenManus)2025.12 | 8,453 | 9,711 | 10,514 | 5,220 | 8,888 | |
| Extra TaskAgent=Cloud-Hosted (OpenManus)2025.12 | 8,433 | 4,065 | 5,768 | 4,507 | 8,831 | |
| Extra TaskAgent=Hybrid (Pydantic-AI)2025.12 | 6,187 | 4,106 | 7,378 | 4,223 | 5,873 |