Spectral theory lecture test: Claude Opus 5 beats ChatGPT Astra at math animations
PTenigma · x · 2026-09-14
PTenigma compared Claude (Opus 5 High) and ChatGPT (Astra Medium) on animations for a spectral theory lecture, slowing them down so the formulas are readable.
- Verdict: Claude's animations (top) came out clearly better.
- Curious inversion: ChatGPT Astra's static math figures are fairly terrible, yet it still produces decent animations of the same content — the author speculates it uses vector graphics for animation but possibly a diffusion transformer (DiT) for still images, which struggles with symbols and lettering.
Related event: Claude Opus 5 beats ChatGPT Astra in math animation test(2 posts)→
More from Models
- Alexandr Wang: Meta spent a year secretly tuning its models for the Muse personal agent — alexandr_wang · 2026-09-14
- GPT-6 Astra: 10 wild demos from game-building to autonomous multi-day work — minchoi · 2026-09-14
- Scale CEO Alexandr Wang backs Meta Muse as shifting economic surplus to consumers — alexandr_wang · 2026-09-14
- Claude Fable 5.1 cracks a 370-year-old cipher in 44 minutes, no human help — bcherny · 2026-09-14
- Next Token Podcast Ep.2: GPT-6 Astra, DeepSeek V4.1 Flash and iPhone Duo — op7418 · 2026-09-14
- Three weird image prompts — pelican on a bike, Wukong flying — to detect model dumbing — lxfater · 2026-09-14