Pipeline Math Project 

The Pipeline Math project aims at understanding:

1. What theoretical problems frontier models are able to solve? How can AI help scientists push the frontiers in theory?

2. Can researchers not in frontier labs do frontier theoretical research using AI?

As of today (July 2nd 2026), our agentic pipeline has solved:

1. Erdős Problem 477

2. COLT open problems (x 4)

3. Commutative algebra open problems from Cahen–Fontana–Frisch–Glaz (x 4)

See our github project link for the solutions. We plan to write a report on our pipeline soon, stay tuned!