Calibration
Artificial Analysis Intelligence Index → GPQA Diamond
GPQA Diamond is estimated from Artificial Analysis Intelligence Index with a offset logistic curve fitted on 43 models measured on both: y = 0.0000 + (0.9309 − 0.0000) / (1 + exp(−11.77·(x − 0.0066))), R² = 0.88, cross-validated error 2.4 pp. It is used for 3 estimates.
| Estimated model | Artificial Analysis Intelligence Index | GPQA Diamond | Source |
|---|---|---|---|
| Claude 4.1 Opus | 18.6% | 83.0% | estimated ± 2.4 pp, high confidence |
| o1-preview | 11.4% | 72.6% | estimated ± 2.4 pp, high confidence |
| o3-pro | 21.9% | 86.0% | estimated ± 2.4 pp, high confidence |