Primary evidence
- FrontierMath: A benchmark for evaluating advanced mathematical reasoning in AIS1 · Primary research · Benchmark · 2024-11-08
FrontierMath introduces expert-crafted research-level mathematics evaluation
The benchmark was developed with more than sixty mathematicians and emphasized problems requiring hours or days of expert work.
What happened?
FrontierMath launched with hundreds of original expert-written problems across modern mathematics; leading models initially solved fewer than 2%.
Why it matters
It created a visible bridge between benchmarked model reasoning and the difficulty regime of genuine mathematical research.
Significance rationale: FrontierMath shifted informal mathematical evaluation from contest problems toward expert-authored problems spanning advanced and research mathematics.
Historical context
Earlier developments
Later developments
—
Related milestones
—
Evidence & status at a glance
- Significance
- H2 · Field milestone
- Verification
- Claimed
- Evidence level
- E3 · Inspectable artifact
- AI role
- Computation only
- Formal assurance
- Artifact available
Contribution & verification
- Mathematical novelty
- Tooling only
- Interface
- Informal, Symbolic, Numeric, Hybrid
- AI contribution
- Frontier models are evaluated with extended reasoning and computational tools on expert-created problems.
- Human contribution
- Mathematicians authored and vetted problems while Epoch AI designed the evaluation infrastructure.
- Verification
- Claimed · E3 · Inspectable artifact · Artifact available
- Last reviewed
- 2026-09-11
Technical details
- Systems
- frontiermath
- Organizations
- epoch-ai
- Problems
- research-mathematics
- Methods
- benchmarking, expert-curation
Sources & artifacts
- FrontierMath: A benchmark for evaluating advanced mathematical reasoning in AIS1 · Primary research · Benchmark · accessed 2026-09-10
- FrontierMath benchmark surfaceArtifact · Dataset
Verification history
- 2024-11-08 · Claimed · E2 · Research record
The Chronicle records this event from the cited primary or official source; the status does not imply independent Chronicle verification.
- 2026-09-11 · Claimed · E3 · Inspectable artifact
This corpus refresh added a directly inspectable public artifact maintained by the authors or releasing organization. E3 records artifact-level inspectability and does not imply independent verification of the underlying scientific or mathematical claim.
Corrections
No corrections recorded.