FrontierMath Tier 4 2025-07-01 Private: leaderboard
Private Tier 4 FrontierMath problems at research-level mathematical difficulty.
Metric: Score (%). Source: neosignal.io. Source: benchmarklist.com. Saturation forecast: Around January 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Sol | 75.7 | #19 |
| 2 | GPT-5.5 | 71.5 | #31 |
| 3 | GPT-5.6 Terra | 68.5 | #41 |
| 4 | GPT-5.6 Luna | 65.2 | #68 |
| 5 | Claude Opus 4.8 | 50 | #37 |
| 6 | Gemini 3.1 Pro (Preview) | 39.1 | #58 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=frontiermath-tier-4-2025-07-01-private · How It Works · Data refreshed daily, snapshot 2026-10-20.