Calibration
HLE w/o tools → AA-HLE
AA-HLE is estimated from HLE w/o tools with a linear curve fitted on 28 models measured on both: y = 0.8049·x + 0.0768, R² = 0.87, cross-validated error 5.0 pp. It is used for 2 estimates.
| Estimated model | HLE w/o tools | AA-HLE | Source |
|---|---|---|---|
| Claude Mythos 5 | 59.0% | 55.2% | estimated ± 5.0 pp, high confidence |
| Pareto 26.9 | 49.0% | 47.1% | estimated ± 5.0 pp, high confidence |