Calibration
SuperGPQA → Artificial Analysis Intelligence Index
Artificial Analysis Intelligence Index is estimated from SuperGPQA with a Michaelis–Menten curve fitted on 16 models measured on both: y = 0.7420·x / (1.55105 + x), R² = 0.43, cross-validated error 3.9 pp. It is used for 4 estimates.
| Estimated model | SuperGPQA | Artificial Analysis Intelligence Index | Source |
|---|---|---|---|
| MiniCPM5-1B | 23.1% | 9.6% | estimated ± 3.9 pp, low confidence |
| Qwen3 235B 2507 | 62.6% | 21.3% | estimated ± 3.9 pp, medium confidence |
| Seed 2.1 Pro | 70.8% | 23.3% | estimated ± 3.9 pp, medium confidence |
| Seed 2.1 Turbo | 67.4% | 22.5% | estimated ± 3.9 pp, medium confidence |