Benchmarks · math
FrontierMath Erdős
Epoch AI's 68 open Erdős problems, formulated in Lean. Models must supply complete formal proofs or disproofs. Published solved fraction; distinct from FrontierMath difficulty tiers. Data: Epoch AI, CC BY.
13 models · 13 results · data as of 2026-10-29 · epoch.ai ↗
Leaderboard
GPT-6.1 Sol max
All results
13 rows| # | Model | Effort | FrontierMath Erdős | Harness |
|---|---|---|---|---|
| 1 | GPT-6.1 Sol | max | 4.4% | Epoch AI |
| 2 | GPT-6 Astra | max | 2.9% | Epoch AI |
| 3 | Opus 5.5 | max | 2.9% | Epoch AI |
| 4 | Sonnet 5.5 | max | 2.9% | Epoch AI |
| 5 | GPT-6 Sol | max | 1.5% | Epoch AI |
| 6 | Fable 5.1 | max | 0.0% | Epoch AI |
| 7 | Fable 5 | max | 0.0% | Epoch AI |
| 8 | Opus 5 | max | 0.0% | Epoch AI |
| 9 | Gemini 3.8 Flash | high | 0.0% | Epoch AI |
| 10 | GPT-5.5 | xhigh | 0.0% | Epoch AI |
| 11 | GPT-5.6 Sol | max | 0.0% | Epoch AI |
| 12 | Grok 4.7 | xhigh | 0.0% | Epoch AI |
| 13 | Kimi K3 | max | 0.0% | Epoch AI |