OpenAI says one model resolved more than 100 open maths problems.
OpenAI said on Monday that a model it began training on 28 August resolved over 100 open problems. It also set up a nine-member advisory group of mathematicians at Princeton's Institute for Advanced Study. The group will judge how results are released, but cannot slow the company's maths work.
Why it mattersTwenty-five winners of maths's top prize had just accused AI labs of trampling their field. OpenAI's answer decides who checks machine proofs before the rest of us see them.
▶ Related video: An AI 'Solved' a Famous Math Problem - Then 9 Mathematicians Checked Its Work
Read in the app — free, in 9 languages
Related stories
Chatbots got money questions wrong more than half the time.2026-09-21Three researchers used Claude to break into OpenAI staff accounts.
2026-09-21Terence Tao is building maths models no AI company controls.
2026-09-20A web developer used AI agents to settle a 50-year-old conjecture.
2026-09-20Zhipu's model tripled the speed of the system that runs it.
2026-09-19