AI Science & Discovery
OpenAI Astra Solves 10 Unsolved Math Problems
The next-generation model uses formal Lean proofs to resolve long-standing challenges in computer science.
A technical illustration featuring complex geometric spheres and mathematical formulas representing an AI breakthrough in solving mathematical problems.
Photo: Kronos News
OpenAI announced on August 1, 2026, that its internal model, Astra, solved 10 previously open mathematical problems [1][2]. These breakthroughs include proving the existence of non-sofic groups and defining new sphere-packing bounds [1]. The achievements mark a significant shift for AI in theoretical computer science [2]. Each solution reportedly required $2,000 in compute resources to complete [1].
To ensure accuracy, OpenAI published formal Lean proofs for the results on GitHub [1]. This method allows researchers to verify the complex logic through computational systems [2]. Experts suggest this demonstrates the potential for agentic AI to handle advanced scientific reasoning [2]. The development signals a new era for automated discovery in high-level mathematics [1].
Editorial notes
Transparency note
AI assisted drafting. Human edited and reviewed.
- AI assisted
- Yes
- Human review
- Yes
- Last updated
Risk assessment
The risk level is set to high because the SOURCE_LIST contains only two independent domains, which is below the recommended minimum of three.
Sources
Related stories
View all
Topics
About the author
Kronos News Desk covers ai science & discovery and editorial analysis for Kronos News.