CaptchaArena: A Large-Scale, Fine-Grained Dataset for Training Computer-Use Agents on Interactive CAPTCHAs
Interactive CAPTCHAs remain challenging for computer-use agents, while existing datasets face trade-offs among type coverage, interaction fidelity, and trajectory supervision. To address these gaps, we present CaptchaArena, the first large-scale, fine-grained training dataset for interactive CAPTCHA solving. It contains 50K puzzles across 20 CAPTCHA types and 5 interaction modes, with every solution verified through execution. CaptchaArena provides 50K screenshot-action trajectories, including 46K with step-by-step reasoning annotations. It also includes fine-grained pixel-mask annotations for irregular targets. Using CaptchaArena, we train CaptchaAgent, a single 9B policy for all 20 CAPT...
| Authors | Zhenhao Zhang, Zhaoyu Fan, Haohan Ying et al. |
|---|---|
| Primary category | cs.CV |
| Published | 25 September 2026 |
| Last updated at source | 25 September 2026 |
Connected in the knowledge graph
- ZHEN-04/CaptchaArenaDataset · cited by
- ZHEN-04/CaptchaArena-TrajectoriesDataset · cited by
Models and datasets referencing this paper
- ZHEN-04/CaptchaArenaDataset
- ZHEN-04/CaptchaArena-TrajectoriesDataset
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.