math benchmark
FrontierMath
A benchmark of hundreds of original, exceptionally challenging mathematics problems crafted and vetted by expert mathematicians, covering most major branches of modern mathematics from number theory and real analysis to algebraic geometry and category theory.
Updated Aug 7, 2026
Published models17
Registry coverage17
MetricScore
EvidenceB
FrontierMath leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | OP | 89.0% | 100.0% | 17 | C | |
| 2 | OP | 84.9% | 93.8% | 17 | C | |
| 3 | OP | 78.6% | 87.5% | 17 | C | |
| 4 | OP | 47.6% | 81.3% | 17 | C | |
| 5 | OP | 40.3% | 75.0% | 17 | C | |
| 6 | OP | 39.6% | 68.8% | 17 | C | |
| 7 | OP | 35.4% | 62.5% | 17 | C | |
| 8 | OP | 26.7% | 56.3% | 17 | C | |
| 9 | OP | 26.7% | 50.0% | 17 | C | |
| 10 | OP | 26.7% | 43.8% | 17 | C | |
| 11 | OP | 26.3% | 37.5% | 17 | C | |
| 12 | OP | 22.1% | 31.3% | 17 | C | |
| 13 | OP | 15.8% | 25.0% | 17 | C | |
| 14 | OP | 9.6% | 18.8% | 17 | C | |
| 15 | OP | 9.2% | 12.5% | 17 | C | |
| 16 | MI | 6.3% | 6.3% | 17 | C | |
| 17 | OP | 5.5% | 0.0% | 17 | C |
Score distribution
Top published rows on the benchmark's original scale.
FrontierMath highlights
The top published results on this benchmark's own scale.
What is FrontierMath?
Definition and scoring fields from the benchmark registry.
A benchmark of hundreds of original, exceptionally challenging mathematics problems crafted and vetted by expert mathematicians, covering most major branches of modern mathematics from number theory and real analysis to algebraic geometry and category theory.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- FrontierMath
- Modality
- text
- Primary category
- math
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- frontiermath|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about FrontierMath.
Which model scores highest on FrontierMath?
GPT-5.6 Sol is currently ranked first with 89.0%.
What does FrontierMath measure?
A benchmark of hundreds of original, exceptionally challenging mathematics problems crafted and vetted by expert mathematicians, covering most major branches of modern mathematics from number theory and real analysis to algebraic geometry and category theory.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
17 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is marked as eligible for the current LLMBoard capability methodology.