Terence Tao's list of AI contributions to Erdős problems (github.com)

In a groundbreaking effort, AI tools, particularly models like AlphaEvolve and ChatGPT, have made substantial contributions to solving various Erdős problems, a collection of unsolved mathematical challenges posed by the famous mathematician Paul Erdős. Recent advances reveal that while some problems saw partial solutions or slight improvements, others yielded full solutions previously considered elusive, marking a significant milestone in mathematical research. Notably, on December 25, 2025, ChatGPT 5.2 Pro provided a complete solution to a problem first addressed by Erdős and Newman in 1977. This development is significant for the AI/ML community as it showcases the growing potential of AI in tackling complex mathematical inquiries, leveraging machine learning’s ability to analyze vast datasets and uncover patterns beyond human capability. The collaboration between renowned mathematicians and AI systems demonstrates a new paradigm where AI not only assists in the mathematical process but also generates valid solutions, potentially reshaping how mathematical research is conducted and accelerating progress in the field. The implications of these findings may extend to various domains, suggesting that AI could play a crucial role in resolving other open problems within mathematics and beyond.