ainewsblitz.com

Breaking

OpenAI Says Internal Model 'Astra' Cracked 10 Open Math Problems, With Proofs Verified in Lean

OpenAI Says Internal Model 'Astra' Cracked 10 Open Math Problems, With Proofs Verified in Lean
  • Foundation Models
  • Research & Papers
  • Open Source

OpenAI on August 1, 2026 announced that an internal, unreleased model dubbed "Astra" produced new results on ten long-standing open problems in mathematics and theoretical computer science, with each argument formalised into a machine-checkable Lean certificate and released publicly. The company laid out the work in a page titled "Ten advances in mathematics and theoretical computer science," describing a hybrid pipeline in which the AI generated the underlying arguments, human researchers prepared them into manuscripts, and each proof was encoded in Lean 4 so its logical correctness could be verified automatically. The full set of Lean files, along with build instructions, was posted to a GitHub repository at openai/ten-proofs.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.

$20
Read this article
$29/month
Unlimited — all 27,934 articles, the full archive, and comprehension quizzes
Save 72%
$98/year
≈ $8.17/month
Unlimited, billed once a year