Calibration
FrontierMath v2 (Tiers 1-3) → FrontierMath v2 (Tier 4)
FrontierMath v2 (Tier 4) is estimated from FrontierMath v2 (Tiers 1-3) with a inverse Michaelis–Menten curve fitted on 56 models measured on both: y = 0.34536·(x − 0.2178) / (0.2178 + 0.9434 − x), R² = 0.99, cross-validated error 4.2 pp. It is used for 8 estimates.
| Estimated model | FrontierMath v2 (Tiers 1-3) | FrontierMath v2 (Tier 4) | Source |
|---|---|---|---|
| o3 | 33.3% | 4.8% | estimated ± 4.2 pp, high confidence |
| Gemini 3.1 Flash-Lite | 27.7% | 2.3% | estimated ± 4.2 pp, high confidence |
| GPT-4.1 | 6.0% | 0.0% | estimated ± 4.2 pp, medium confidence |
| GPT-4.1 mini | 6.7% | 0.0% | estimated ± 4.2 pp, medium confidence |
| GPT-4o | 0.4% | 0.0% | estimated ± 4.2 pp, medium confidence |
| o1 | 14.7% | 0.0% | estimated ± 4.2 pp, medium confidence |
| Qwen3.5 Flash | 18.2% | 0.0% | estimated ± 4.2 pp, medium confidence |
| Qwen3.7 Flash | 19.3% | 0.0% | estimated ± 4.2 pp, medium confidence |