FrontierMath

Roadmap

Unpublished research-level math problems vetted by professional mathematicians.

On the roadmap

FrontierMath is catalogued but not runnable yet, so there are no usage docs — we do not document what does not run. The fact sheet below is sourced from the paper; the protocols it will implement are stable today.

Paper: FrontierMath: A Benchmark for Evaluating Advanced Mathematical Reasoning
Citation: Glazer et al., 2024, arXiv:2411.04872
License: Proprietary (Epoch AI)
Homepage: https://epoch.ai/frontiermath

How an eval goes live

Implement an EvalRunner against the stable protocols.
Bundle a small real-schema sample so it runs offline.
Point the catalog entry's runner at the class.
Ship its docs in the same change — required to flip live.

pip install agi-evals