Calibration
τ³-bench results → WideResearch
WideResearch is estimated from τ³-bench results with a Hill curve fitted on 7 models measured on both: y = 1.0000 + (0.7039 − 1.0000)·x^6.00 / (0.37906^6.00 + x^6.00), R² = 0.35, cross-validated error 12.4 pp. It is used for 1 estimate.
| Estimated model | τ³-bench results | WideResearch | Source |
|---|---|---|---|
| Pokee-Isaac 28B | 66.2% | 71.4% | estimated ± 12.4 pp, low confidence |