// Design Challenge
Same brief, several models: what does each one build?
They all get the same prompt. Each one builds its site, and we publish it as delivered and working, so you can compare them and try them yourself.
How we run it
-
The same prompt for everyone
If a model got an extra instruction, we say so on its page.
-
The sites, as delivered
We serve them without touching their code. The only addition is a few lines so scrolling can sync when you compare them.
-
We measure, then we judge
A script measures weight, errors, overflow, motion and accessibility. The design score is a critique: an opinion, with its reasons.
Looking for something more measurable? MondayBench tests models on real work