Benchmark RadarAI BENCHMARK PROFILE
DragOn
DragOn is a drag grounding benchmark and training dataset for GUI agents across four interaction types: text highlighting, cell selection, element resizing, and slider manipulation, with 3.5M training tasks and 2,000 evaluation examples.
- Released
- 2026-06-04
- Readiness
- Paper only
- Primary field
- Robotics & Autonomous Systems
Why it matters
Drag-based interactions are under-represented in existing datasets. DragOn provides a large-scale resource for improving drag grounding, which may enhance performance on downstream computer-use tasks.
Motivation
GUI agents - vision-based models that control desktops, web browsers, and mobile devices through graphical user interfaces - promise to automate a wide range of digital tasks.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.