WebsiteBench

Can AI agents rebuild websites through browser exploration?

About WebsiteBench

WebsiteBench is a benchmark of AI agents rebuilding 70 production-level websites through browser exploration: 46,916 hidden tests, eight models, best score 26.2%.

Ratings & reviews

–0 ratings
5★0
4★0
3★0
2★0
1★0

Log in to rate and review WebsiteBench.

Discussion (0)

Log in to join the discussion.