Report: OpenAI's Internal Model Solves 10+ Open Math Problems, AI-Assisted Proofs Gain Traction

jiaxinwen22 · x · 2026-08-09

Theoretical computer scientist Scott Aaronson revealed on his blog that an internal OpenAI model has solved ten significant open problems in math and theoretical computer science. These include parallel repetition for arbitrary quantum games and polynomial-factor hardness of approximation for the Closest Vector Problem (CVP).

Separately, Frederic Koehler and Pui Kuen Leung announced a proof for the Permanent Anti-Concentration Conjecture proposed by Aaronson 16 years ago, acknowledging that they discussed ideas with ChatGPT. Aaronson highlighted that multiple AIs are now breaking through barriers in solving complex mathematical problems.

Related event: Frontier AI Models Solve 25-Year-Old Wireless Communication Problem(12 posts)→

Original post →

More from AGI Musings

AGI Musings channel →