ResearchBenchmarksGUI Agent Task Success on OSWorld VLC (test)Follow70.7Success RateGPT-4o52.70857.37962.0566.721Jun 2, 2026Evaluation ResultsMethodMethodLinksSuccess RateGPT-4oModel Setup=+TutorialModel Setup=+Tutorial2026.0670.7GPT-4oModel Setup=baselineModel Setup=baseline2026.0659.6o1-miniModel Setup=+TutorialModel Setup=+Tutorial2026.0656.1o1-miniModel Setup=baselineModel Setup=baseline2026.0653.4