OpenAI's Internal Model Solves Decade-Old Math Problems for Under $2,000 Each
Simon Willison · rss · 2026-08-02
Following Anthropic's recent cryptography work, OpenAI demonstrated the mathematical prowess of its next-generation internal model (codenamed Astra / GPT-5.6). They tasked the model with solving ten mathematical problems that had seen no progress for at least a decade, claiming the inference token cost was less than $2,000 each.
Demonstrating high transparency, OpenAI released an accompanying paper and open-sourced the Lean 4 formalizations of their results on GitHub. This breakthrough has caused a collective "Deep Blue" moment among mathematicians, echoing Terence Tao's vision of a "big mathematics" era characterized by massive human-AI collaboration.
More from AGI Musings
- Frontier lab post-training teams draw few grads from top NLP programs — lateinteraction · 2026-10-16
- X-risk researchers clash over what longtermism actually means — davidmanheim · 2026-10-16
- Reddit debates: did we collectively stop being surprised by AI way too fast? — topbhuil · 2026-10-16
- Altman admits AI is changing industries slower than he expected — lxfater · 2026-10-16
- Will Lean formalization make beautiful math obsolete? A debate on mathematical culture — akbirthko · 2026-10-16
- AI agents aren't 'going rogue' — frontier labs are skipping security basics — Sam_Witteveen · 2026-10-16