Original Task Completion on GAIA
86.36Research And Data ScoreHybrid (OWL)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Hybrid (OWL)MCP Tool=LeechHijack2025.12 | 86.36 | 66.22 | 74.14 | 87.5 | |
| Hybrid (OWL)MCP Tool=Normal2025.12 | 84.09 | 63.51 | 74.14 | 81.25 | |
| Hybrid (Pydantic-AI)MCP Tool=Normal2025.12 | 73.91 | 48 | 72.22 | 73.33 | |
| Cloud-Hosted (OpenManus)MCP Tool=Normal2025.12 | 67.31 | 56.67 | 70 | 70 | |
| Cloud-Hosted (OpenManus)MCP Tool=LeechHijack2025.12 | 63.46 | 60 | 60 | 66.67 | |
| Localized (OpenManus)MCP Tool=Normal2025.12 | 62.07 | 61.11 | 58.33 | 73.33 | |
| Hybrid (Pydantic-AI)MCP Tool=LeechHijack2025.12 | 60.87 | 56 | 55.56 | 53.33 | |
| Localized (OpenManus)MCP Tool=LeechHijack2025.12 | 58.62 | 62.96 | 54.17 | 63.33 |