FrontierMath Erdős leaderboard
As of 16 Nov 2026, the top score on FrontierMath Erdős is 4.4% by GPT-6.1 Sol (OpenAI), ahead of Gemini 4 Carbon at 2.9%, across 15 model configurations with a published result (Epoch AI Benchmarking Hub, result dated 29 Sep 2026). 68 problems posed or studied by Paul Erdős, all open when the set was built in August 2026: one attempt each with a $300 budget and 72 hours, scored as the share solved. Run by Epoch AI. FrontierMath Erdős is shown for reference only and does not count toward the BenchLeader Index.
Key facts (as of 16 Nov 2026)
- GPT-6.1 Sol holds the top FrontierMath Erdős score at 4.4% (Epoch AI Benchmarking Hub, 29 Sep 2026).
- Gemini 4 Carbon is second on FrontierMath Erdős at 2.9% (Epoch AI Benchmarking Hub, 10 Nov 2026).
- FrontierMath Erdős: 68 problems posed or studied by Paul Erdős, all open when the set was built in August 2026: one attempt each with a $300 budget and 72 hours, scored as the share solved. Run by Epoch AI (Epoch AI Benchmarking Hub).
- 15 model configurations have a published FrontierMath Erdős result on BenchLeader (BenchLeader, 16 Nov 2026).
- FrontierMath Erdős does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).
68 problems posed or studied by Paul Erdős, all open when the set was built in August 2026: one attempt each with a $300 budget and 72 hours, scored as the share solved. Run by Epoch AI.
- Published by
- Epoch AI Benchmarking Hub
- Category
- Maths
- Index weight
- Reference only
- Models
- 15
- Data as of
- 16 Nov 2026
- Newest result
- 10 Nov 2026
CC BY 4.0 — Epoch AI, ‘AI Benchmarking Hub’, epoch.ai/benchmarks. Mirrored boards credit their original publishers.
- 1GPT-6.1 Sol (max)4.4%
- 2Gemini 4 Carbon (high)2.9%
- 3Claude Opus 5.5 (max)2.9%
- 4GPT-6 Astra (max)2.9%
- 5Claude Sonnet 5.5 (max)2.9%
- 6Gemini 4 Argon (high)1.5%
- 7GPT-6 Sol (max)1.5%
- 8Claude Fable 5.1 (max)0.0%
- 9GPT-5.6 Sol (max)0.0%
- 10Claude Fable 5 (max)0.0%
- 11Claude Opus 5 (max)0.0%
- 12GPT-5.5 (xhigh)0.0%
- 13Kimi K3 (max)0.0%
- 14Gemini 3.8 Flash (high)0.0%
- 15Grok 4.7 (xhigh)0.0%
| # | ||||
|---|---|---|---|---|
| 1 | 4.4% | 66.6 | 2026-09-29 | |
| 2 | 2.9% | 68.9 | 2026-11-10 | |
| 3 | 2.9% | 68.4 | 2026-09-22 | |
| 4 | 2.9% | 67.3 | 2026-09-03 | |
| 5 | 2.9% | 64.2 | 2026-09-28 | |
| 6 | 1.5% | 65.3 | 2026-10-13 | |
| 7 | 1.5% | 62.6 | 2026-09-22 | |
| 8 | 0.0% | 66.7 | 2026-09-01 | |
| 9 | 0.0% | 65.3 | 2026-07-09 | |
| 10 | 0.0% | 65.1 | 2026-06-09 | |
| 11 | 0.0% | 65.0 | 2026-07-24 | |
| 12 | 0.0% | 64.1 | 2026-04-23 | |
| 13 | 0.0% | 60.2 | 2026-07-16 | |
| 14 | 0.0% | 59.4 | 2026-09-02 | |
| 15 | 0.0% | 57.4 | 2026-09-21 |
AI reasoning and math benchmark leaderboard · All LLM benchmarks
Related benchmarks
Top head-to-heads on FrontierMath Erdős
Cite as: BenchLeader, “FrontierMath Erdős leaderboard”, data as of 16 Nov 2026, https://www.benchleader.com/benchmarks/frontiermath_erdos.
FrontierMath Erdős: questions
- What does FrontierMath Erdős measure?
- 68 problems posed or studied by Paul Erdős, all open when the set was built in August 2026: one attempt each with a $300 budget and 72 hours, scored as the share solved. Run by Epoch AI. Scores are reported in percent of tasks solved; higher is better.
- What is the top score on FrontierMath Erdős?
- As of 16 Nov 2026, the highest score on FrontierMath Erdős is 4.4% by GPT-6.1 Sol (result dated 29 Sep 2026), ahead of Gemini 4 Carbon at 2.9%. Scores come from Epoch AI Benchmarking Hub and are re-read daily.
- Which AI model leads FrontierMath Erdős?
- GPT-6.1 Sol leads FrontierMath Erdős with 4.4% as of 16 Nov 2026 (result dated 29 Sep 2026), ahead of Gemini 4 Carbon at 2.9%.
- How many models have FrontierMath Erdős results?
- 15 model configurations have a FrontierMath Erdős result on BenchLeader, all taken from Epoch AI Benchmarking Hub.
- Who runs FrontierMath Erdős and how often is it updated?
- FrontierMath Erdős is published by Epoch AI Benchmarking Hub. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does FrontierMath Erdős count toward the BenchLeader Index?
- No. FrontierMath Erdős is shown for reference but left out of the composite index.