Calibration
ARC-AGI-3 → HealthBench Professional (raw)
HealthBench Professional (raw) is estimated from ARC-AGI-3 with a linear curve fitted on 5 models measured on both: y = 0.1309·x + 0.6197, R² = 0.42, cross-validated error 5.3 pp. It is used for 1 estimate.
| Estimated model | ARC-AGI-3 | HealthBench Professional (raw) | Source |
|---|---|---|---|
| Grok 4.20 | 0.1% | 62.0% | estimated ± 5.3 pp, low confidence |