FrontierMath Tiers 1-3 (v2) - Private
Math · 2026-06-12
The Private split of FrontierMath Tiers 1-3 (v2). This child benchmark separates a source-reported subtask or subtrack from the parent aggregate so scores at different grains do not share one benchmark_slug.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-5.6 Sol | 89.1 |
| GPT-5.5 Pro | 87.7 |
| Claude Fable 5 | 87.0 |
| GPT-5.6 Terra | 86.0 |
| Claude Opus 5 | 85.6 |
| GPT-5.5 | 85.3 |
| GPT-5.4 Pro | 82.5 |
| GPT-5.6 Luna | 82.1 |
| Opus 4.8 | 80.0 |
| GPT-5.4 | 78.6 |
| GPT-5.2 Pro | 74.0 |
| Kimi K3 | 72.2 |
| Opus 4.7 | 70.2 |
| GPT-5.2 | 67.4 |
| Opus 4.6 | 66.0 |
| Sonnet 5 | 65.6 |
| Qwen3.7-Max | 64.6 |
| Gemini 3.5 Flash | 62.8 |
| GLM-5.2 | 59.2 |
| Gemini 3.6 Flash | 58.9 |
| Grok 4.5 | 57.2 |
| Kimi K2.6 | 57.2 |
| GPT-5 Pro | 55.8 |
| GPT-5 | 55.4 |
| Kimi K2.7 Code | 54.0 |
| Gemini 3 Flash Preview | 51.2 |
| GPT-5.4 Mini | 51.2 |
| GPT-5 Mini | 46.7 |
| DeepSeek-V4-Pro | 45.3 |
| GPT-5.4 Nano | 44.9 |
| Grok 4.20 | 44.9 |
| Grok 4.3 | 42.8 |
| O4 Mini | 36.1 |
| Opus 4.5 | 34.4 |
| Gemini 2.5 Pro | 24.6 |
| Sonnet 4.5 | 23.9 |
| GPT-5 Nano | 20.0 |
| o3-mini | 18.6 |
| Opus 4.1 | 12.6 |