FrontierMath
Several hundred unpublished, research-level maths problems; v2 reports Tiers 1-3 and Tier 4 separately; private Tier 4
Tiers 1-4mathactive
Definition
- Version
- Tiers 1-4
- Category
- math
- Maintainer
- Epoch AI
- Status
- active
- Home page
- https://epoch.ai/frontiermath
Sources
- FrontierMath: A Benchmark for Evaluating Advanced Mathematical Reasoning in AI arxiv.org https://arxiv.org/abs/2411.04872
- FrontierMath (legacy) Leaderboard: GPT-5.6 Sol Leads at 89% benchlm.ai https://www.benchlm.ai/benchmarks/frontiermath
- AI benchmarks & capabilities epoch.ai https://epoch.ai/benchmarks
Scores
| Model | Score | Who ran it | Setting | View or population | Date | Source |
|---|---|---|---|---|---|---|
 GPT-5.6 Sol | 89 % | vendor-reported | Legacy aggregate mixing benchmark versions; self-report | rank 1, BenchLM aggregator, top score | 10 Oct 2026 | benchlm.ai |
Where this comes from
Scores from different settings, harnesses or dates are not directly comparable; the row says which. Sorted by score, highest first, as a reading aid, not a ranking.
To cite this page: Project AURION, AI Use-Cases Research & Innovation, "FrontierMath", https://aurion.semai.id/benchmarks/frontiermath/, records of October 2026.