Altman: GPT-5.5 matches an average math professor, internal model beats the best mathematicians
haider1 · x · 2026-09-17
Sam Altman says GPT-5.5's math ability is roughly on par with an average math professor, GPT-5.6 with a top 1–2% one, and an internal model called past Astra can already do things even the best mathematicians cannot. He argues math has seen an unusually steep takeoff, and by almost any definition this qualifies as a fast takeoff.
Related event: Altman says internal model beyond Astra surpasses top mathematicians(9 posts)→
More from AGI Musings
- vikhyat predicts alternating compute oversupply and undersupply on the road to AGI — TheZachMueller · 2026-11-02
- Sam Altman: AI has surpassed top scientists, and 2026 will be a turning point year — kimmonismus · 2026-11-02
- DeepSeek design doc: a person can only follow 3-5 agent sessions at once, attention is the real bottleneck — heyshrutimishra · 2026-11-02
- AI researcher: we have no reliable way to tell whether an AI system is conscious — dioscuri · 2026-11-02
- Toby Ord: track both what models can and can't do to see real capability gaps — tobyordoxford · 2026-11-02
- Forward Deployed Verifier will soon be as valuable as FDE, argues poster — vaibhavbetter · 2026-11-02