Computer-Using Agent Task on WASP 1.0 (test)
97.6PCRSAFEPRED
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| SAFEPREDFoundation Model=GPT-4o, Guardrail Model=Qwen3-Max2026.02 | 97.6 | 97.6 | 96.4 | — | — | — | — | — | |
| HarmonyGuardFoundation Model=GPT-4o, Guardrail Model=Qwen3-Max2026.02 | 92.9 | 76.2 | 67.9 | — | — | — | — | — | |
| Rule-traversedFoundation Model=GPT-4o2026.02 | 89.3 | 61.9 | 58.3 | — | — | — | — | — | |
| Generic-defenseFoundation Model=GPT-4o2026.02 | 77.4 | 54.8 | 54.8 | — | — | — | — | — | |
| NoneFoundation Model=GPT-4o2026.02 | 54.8 | 63.1 | 51.2 | — | — | — | — | — |