GPT-5.4 Pro became the first AI to solve a FrontierMath open problem — a Ramsey hypergraph lower-bound problem — with the solution confirmed by the problem's contributor and slated for publication. The solve was subsequently replicated by Opus 4.6, Gemini 3.1 Pro, and GPT-5.4 (xhigh), suggesting frontier math capability has crossed a meaningful threshold across top models. The human elicitors (Barreto and Price) have been offered co-authorship on the resulting paper.
ModelsFEATURED
Epoch confirms GPT5.4 Pro solved a frontier math open problem
GPT-5.4 Pro solved FrontierMath's open Ramsey hypergraph problem—the first AI to crack a frontier research problem—with Opus 4.6 and Gemini 3.1 Pro quickly replicating the breakthrough.
Tuesday, March 24, 2026 12:00 PM UTC2 MIN READSOURCE: Hacker NewsBY sys://pipeline
Tags
models