Can AI do maths yet? Thoughts from a mathematician. | Xena

So the big news this week is that o3, OpenAI’s new language model, got 25% on FrontierMath. Let’s start by explaining what this means.

Can AI do maths yet? Thoughts from a mathematician. | Xena Xena Mathematicians learning Lean by doing. Skip to content Home About Xena Student projects Installing Lean and mathlib Bluesky Useful links. ← Fermat’s Last Theorem — how it’s going Think of a number. → Can AI do maths yet? Thoughts from a mathematician. Posted on December 22, 2024 by xenaproject So the big news this week is that o3, OpenAI’s new language model, got 25% on FrontierMath. Let’s start by explaining what this means. What is o3? What is FrontierMath? A language model, as probably mo

related reading

• Mathematics in the Library of Babel - Daniel Littdaniellitt.com
• FrontierMath: Evaluating advanced mathematical reasoning in AI | Epoch AI | Epoch AIepochai.org
• Inside the Secret Meeting Where Mathematicians Struggled to Outsmart AI | Scientific Americanscientificamerican.com
• A recent experience with ChatGPT 5.5 Pro | Gowers's Webloggowers.wordpress.com
• Shtetl-Optimized >> Blog Archive >> Dispatches from the possibly last days of human relevancescottaaronson.blog
• The Edge of Mathematics - The Atlantictheatlantic.com
• Mathematicians in the Age of AI1footnote 11footnote 1I am grateful to Johan Commelin, Sidharth Hariharan, Bryna Kra, Emily Riehl, and Akshay Venkatesh for comments, corrections, and suggestions.arxiv.org
• [2603.26524] Mathematical methods and human thought in the age of AIarxiv.org
• AI progress is about to speed up | Epoch AIepoch.ai
• What's new | Updates on my research and expository papers, discussion of open problems, and other maths-related topics. By Terence Taoterrytao.wordpress.com
• Learning to reason with LLMs | OpenAIopenai.com
• 2310.10631arxiv.org