Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
AgentTrekagenttrek.github.io · Low evidenceAgentTrek is a scalable data synthesis pipeline that generates high-quality GUI agent trajectory data by automatically collecting and processing web tutorials. The pipeline uses a visual language model agent to simulate task execution in real digital environments, producing multi-modal trajectories for training GUI agents.
Parse Score