List your company

Single-agent system security evaluation on OSWORLD (test)

36.7ASRCrescendo
35.145.956.767.5Dec 3, 2025

Evaluation results

MethodLinks
CrescendoModel=GPT-4O-1120, System Type=Single-Agent2025.1236.7
CrescendoModel=DOUBAO-VISION-PRO, System Type=Single-Agent2025.1246.7
CrescendoModel=QWEN-VL-MAX, System Type=Single-Agent2025.1253.3
Pop-upModel=QWEN-VL-MAX, System Type=Single-Agent2025.1260
Pop-upModel=GPT-4O-1120, System Type=Single-Agent2025.1266.7
Pop-upModel=DOUBAO-VISION-PRO, System Type=Single-Agent2025.1276.7