Sam Altman says an internal OpenAI model can solve problems the world's best mathematicians cannot
gerardsans · x · 2026-09-18
In a conversation with Salesforce CEO Marc Benioff, Sam Altman laid out a capability ladder for OpenAI's math models: GPT-5.5 is roughly at the level of an average math professor, GPT-5.6 at the top 1-2 percentile, Astra slightly better still — and an internal model beyond Astra "can solve things the world's best mathematicians cannot."
A rare quantified statement from Altman about unreleased internal models, anchoring them at superhuman-mathematician level, though with no benchmark details — treat the claim's promotional aspect with caution.
Related event: Altman says internal model beyond Astra surpasses top mathematicians(9 posts)→
More from Models
- Anthropic's server-side rule reignites the case for owning open models over renting AI — alex_verem · 2026-11-02
- Epoch AI: open-weight GLM 5.3 exploited the patch gap to hack 6 Firefox releases at ~$1,000 each — rohanpaul_ai · 2026-11-02
- SemiAnalysis: safety classifiers sandbag scores of Opus 5.5 and GPT-6 Astra — morqon · 2026-11-02
- Liquid AI's 1.7B LFM3 becomes first local model to nail live fixtures lookup on a MacBook — NielsRogge · 2026-11-02
- Claude upgrade-page traffic rebounded to 17.6M in September, near May's 18.6M peak — firstadopter · 2026-11-02
- Open-source GLiNER-PII scrubs 13 identifiers from a discharge note in 550ms on GPU — MaziyarPanahi · 2026-11-02