Claude vs ChatGPT math animations: static images terrible, animations surprisingly good

PTenigma · x · 2026-09-14

A lecturer preparing spectral theory lecture animations compared Claude (Opus 5 High) against ChatGPT (Astra Medium): ChatGPT's static math figures were fairly terrible, yet it produced noticeably better animations of the same content.

The author speculates the discrepancy comes from animation being rendered via vector graphics while still images likely rely on a diffusion transformer (DiT), which struggles with symbols and lettering.

Related event: Claude Opus 5 beats ChatGPT Astra in math animation test(2 posts)→

Original post →

More from Models

Models channel →