DragOn: leaderboard
Metric: Success rate (%) on the four drag domains (text highlighting, spreadsheet cell selection, slider manipulation and slide element resizing) of the private 1,000-example DragOn test split (250 per domain): from one screenshot and instruction the model outputs the drag's start and end points as JSON, and both must fall inside the pixel-aligned source and target boxes; each model at its documented settings (Claude extended thinking off, Kimi-K2.5 reasoning off, local Qwen and Holo models with chain-of-thought on); higher is better. Source: arxiv.org. Saturation forecast: Around July 2027. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.7 | 27.7 |
| 2 | GPT-5.4 | 25.7 |
| 3 | Qwen 3.5 397B A17B | 21.5 |
| 4 | Kimi K2.5 (Non-reasoning) | 11.6 |
| 5 | Claude Sonnet 4.5 | 10.7 |
| 6 | Qwen 3.5 35B A3B | 2.3 |
Interactive version: theaggregate.ai/benchmark?slug=dragon · How It Works · Data refreshed daily, snapshot 2026-09-29.