Sam Altman: GPT-6 Astra Could Accelerate Scientific Discovery — Lean Proof of Bounded Prime Gaps

Sam Altman ·

Key Info

Sam Altman highlights GPT-6 Astra's potential for scientific discovery, pointing to an OpenAI repository where the model formalized in Lean a proof that infinitely many pairs of consecutive primes differ by at most 186.

Highlights

  • Altman says scientific discovery is the GPT-6 Astra use case he is most excited about, and the model has not yet been pushed to its limits in math and science.
  • A linked OpenAI repository shows a Lean formalization by GPT-6-Astra of the bounded prime gaps result (gap ≤ 186).
  • The announcement frames the model as a tool for generating new mathematical breakthroughs.
Loading...