ATRIUMsearch → argument graph
ExampleArticle

The model produces substantially different-looking pelicans at low, medium, and high reasoning levels.

The author observes that the pelican images differ markedly depending on the selected reasoning level. They say this degree of variation is unusual compared with other models they have tested. ✦ AI generated

the author · Simon Willison's Weblog · 2026-08-12 · original ↗

Interestingly I got very different looking pelicans for the three different reasoning levels of low, medium, and high. I've not noticed this kind of difference from any other model:

Read full article ↗excerpt · fair-use quotation

Around this claim