Evaluation & Benchmarks · 2024
FrontierMath: A Benchmark for Evaluating Advanced Mathematical Reasoning in AI
Hundreds of original, unpublished research-level math problems vetted by professional mathematicians, on which frontier models initially solved under 2%, exposing the ceiling of math evaluation.
Editorial record
Plain-language summary
Hundreds of original, unpublished research-level math problems vetted by professional mathematicians, on which frontier models initially solved under 2%, exposing the ceiling of math evaluation.
Knowledge graph
Relationships
Descendants
Conceptual ancestorEvidence: Strongly supported
MMLU: Massive Multitask Language Understanding
Builds on the evaluation lineage in the archive
freshness sweep 2026
Source record
Provenance
- Record ID
- P-620
- Record created
- 2026-07-13
- Last reviewed
- 2026-07-14
- Record version
- 2
- https://arxiv.org/abs/2411.04872
- arXiv:2411.04872
Citation caveat: Citation metadata is approximate and marked unverified in the source dataset.