Ask Janek AI with Janek

All questions

I Tested Every Cursor Model on Normal Work · published 1 November 2026

Play at 13:32 Watch on YouTube Cursor

Which Cursor model drew a better picture in code?

Gemini on low was first. Claude was second. ChatGPT was third. Muse was fourth. Grok was last.

The prompt was an SVG, a moose riding a scooter through a forest. That is code drawing a picture, not an image generator. Claude's moose was more realistic. ChatGPT was more cartoon and more playful, and it still looked more professional than Muse. Muse looked like a child's drawing, and the job was still done. Grok had a sun and a playful scene, and the moose itself was the bad part. Gemini on low had one odd patch and was still the one I liked most, including over ChatGPT.