List your company

Trajectory completion judgment on OGRBench Windows

85.5AccuracyOS-Themis
72.70876.02979.3582.671Jun 5, 2026

Evaluation results

MethodLinks
OS-ThemisVerifier=Gemini-3-Flash2026.0685.583.2
StainFlowVerifier=Qwen3.5-27B2026.068583.7
OS-ThemisVerifier=Qwen3.5-27B2026.068481.1
StainFlowVerifier=Gemini-3-Flash2026.068483
ZeroGUIVerifier=Qwen3.5-27B2026.0682.278.9
StainFlowVerifier=Qwen3.5-9B2026.0682.281
DigiRLVerifier=Gemini-3-Flash2026.0681.776.4
DigiRLVerifier=GPT-52026.0680.875.7
ZeroGUIVerifier=GPT-52026.0679.877.5
ZeroGUIVerifier=Gemini-3-Flash2026.0678.977.2
StainFlowVerifier=Qwen3-VL-8B2026.0678.477.2
StainFlowVerifier=GPT-52026.0678.477
DigiRLVerifier=Qwen3.5-27B2026.0677.568.8
DigiRLVerifier=Qwen3-VL-8B2026.067770.7
OS-ThemisVerifier=Qwen3.5-9B2026.067771
OS-ThemisVerifier=GPT-52026.0676.167.5
OS-ThemisVerifier=Qwen3-VL-8B2026.0675.168.3
DigiRLVerifier=Qwen3.5-9B2026.0674.262.1
ZeroGUIVerifier=Qwen3.5-9B2026.0673.771.7
ZeroGUIVerifier=Qwen3-VL-8B2026.0673.274.9