Pelican benchmark
Create an HTML page with an SVG 2D animation of a pelican riding a bicycle.
Loading artwork…
About the pelican benchmark
Every 30 minutes we give each model the same task and show what it produces as-is, without picking or editing, so you can compare how different models write code and draw.
The task: Create an HTML page with an SVG 2D animation of a pelican riding a bicycle.
“A pelican riding a bicycle” is a popular informal test for language models among developers: the task fits in one sentence, the result is easy to judge at a glance, and it is hard for a model to be tuned for it specifically.