GPT-6 Astra claims first improvement on prime gaps since the 1930s, shaving a log log n factor

rapha_gl · x · 2026-09-04

Announced as part of the GPT-6 Astra launch, the model reportedly improved the bound on the longest gap between primes by roughly a log log n factor — the first improvement of that size since the 1930s.

Context from the thread: prior progress by subsets of Ford, Green, Konyagin, Maynard and Tao, and more recently by GPT 5.6 Sol, were smaller log log log n factors. If verified, it would mark another landmark AI-for-math result.

Related event: OpenAI open-sources GPT-6-Astra's Lean proof cutting prime gap bound to 186(16 posts)→

Original post →

More from Models

Models channel →