Calibration
Artificial Analysis Intelligence Index → HLE w/o tools
HLE w/o tools is estimated from Artificial Analysis Intelligence Index with a Michaelis–Menten curve fitted on 28 models measured on both: y = 2.0000·x / (1.37420 + x), R² = 0.83, cross-validated error 6.3 pp. It is used for 3 estimates.
| Estimated model | Artificial Analysis Intelligence Index | HLE w/o tools | Source |
|---|---|---|---|
| Claude 4.1 Opus | 18.6% | 23.8% | estimated ± 6.3 pp, medium confidence |
| o1-preview | 11.4% | 15.3% | estimated ± 6.3 pp, low confidence |
| o3-pro | 21.9% | 27.5% | estimated ± 6.3 pp, medium confidence |