Justin Sun Prize Awards AI‑Assisted Solutions to Two Erdős Problems
The Office of Justin Sun honored three Purdue students and independent researcher for work on Erdős Problem #97 and a Lean formalization of Problem #841, with AI tools playing a key role.
Key takeaways
- Liam Kruer, Jensen Kohlmeyer and Liam Price received the Justin Sun Prize for solving Erdős Problem #97.
- Peilin Liu was awarded for a Lean formalization of Erdős Problem #841.
- Prize payouts can be claimed in USDT on TRON (TRC‑20) or USDC on Ethereum (ERC‑20).
The Office of Justin Sun announced the latest recipients of the Justin Sun Prize on 16 November 2026, recognizing AI‑enhanced research on two long‑standing Erdős problems. Three Purdue affiliates – engineering student Liam Kruer, mechanical‑engineering and mathematics student Jensen Kohlmeyer, and independent researcher Liam Price – were honored for their solution to Erdős Problem #97, while mathematician Peilin Liu received the prize for a Lean formalization of Erdős Problem #841. The awards underscore the growing role of artificial‑intelligence tools such as Codex and ChatGPT Pro in advancing pure mathematics.
Solution to Erdős Problem #97
Problem #97 asks whether every convex polygon contains a vertex from which distances to other vertices are not overly repetitive. According to the prize announcement, Kruer, Kohlmeyer and Price constructed a convex polygon where each vertex has at least four other vertices exactly one unit away, addressing both the original question and a broader variant. The trio contributed equally to the mathematical proof and its formal verification in the Lean theorem prover, while acknowledging substantial AI assistance throughout the process.
Lean Formalization of Problem #841
Erdős Problem #841 concerns the minimal length of a consecutive integer interval that still yields larger integers whose product with a given integer forms a perfect square. Peilin Liu produced a complete Lean formalization of the problem, employing AI tools—including Codex and ChatGPT Pro—to assist in encoding the proof. Liu’s work demonstrates how AI can streamline the translation of informal mathematical arguments into machine‑checkable code.
Prize Mechanics and Pilot Program
The awards are part of a pilot collaboration with MathDB, a research platform created by Caltech professor Tony Yue Yu. The pilot introduces a formal submission and review workflow for AI‑assisted research across fifty designated problems. Winners may receive their prize in either USDT on the TRON network (TRC‑20) or USDC on Ethereum (ERC‑20), subject to compliance verification. All source code, verification records, and further details are hosted publicly on the prize’s GitHub repository.
What to Watch Next
Future prize cycles will expand the list of Erdős challenges and may incorporate additional AI models, offering a view into how decentralized incentives and open‑source verification could reshape mathematical research.


Discussion 0