FrontierMath
A frontier math benchmark with constrained public access and source-linked result claims.
What this benchmark measures
A frontier math benchmark with constrained public access and source-linked result claims.
Rows on this page are sourced from public benchmark artifacts, leaderboard exports, or source-linked model reports. Each row keeps benchmark version, source model name, and available run details attached to the score.
The metric shown here is accuracy. It should be interpreted within FrontierMath, not compared as part of a site-wide ranking.
What to be careful about
Do not include unverifiable scraped claims.
Frequently asked
What is FrontierMath?
A frontier math benchmark with constrained public access and source-linked result claims. It is a reasoning benchmark measured by accuracy.
What does accuracy mean on FrontierMath?
FrontierMath reports accuracy (%); higher is better. Scores are shown only within FrontierMath and are never averaged with other benchmarks.
What is the top reported FrontierMath score?
GPT-5.5 Pro has the top reported score on FrontierMath: 52.4% (accuracy).
Why do FrontierMath scores differ across runs?
Harness, scaffold, reasoning effort, and prompt setup change results, so two runs of the same model can differ. evals.report keeps each score with its run context so the differences stay visible.
Does evals.report rank models across benchmarks?
No. FrontierMath scores are shown within their own metric; evals.report never combines benchmarks into a composite ranking or a single "best model".