Websites · PROJECT NOTES
VisualWebBench
A static multimodal web benchmark covering page understanding, OCR, element localization and action grounding from website screenshots.
What can you do with it?
Identify whether a web agent's visual model struggles with reading, page semantics or locating controls.
What should you know before using it?
Question-answer and grounding evaluation does not execute multi-step browsing tasks. The seven task subsets should be reported separately when diagnosing a model.
What is verified here?
This entry was reviewed against original project material. We have not independently run the software or reproduced its results. AI-development claims are author-reported; check upstream for current versions, setup requirements, and licenses.
- Resource type
- Dataset + benchmark
- Provenance
- Development provenance not established
- Reviewed
- Platforms
- Web