Calibration
FrontierCode 1.1 Main → PostTrainBench v1.1
PostTrainBench v1.1 is estimated from FrontierCode 1.1 Main with a Hill curve fitted on 6 models measured on both: y = 0.2331 + (1.2000 − 0.2331)·x^6.00 / (0.67480^6.00 + x^6.00), R² = 0.76, cross-validated error 5.9 pp. It is used for 2 estimates.
| Estimated model | FrontierCode 1.1 Main | PostTrainBench v1.1 | Source |
|---|---|---|---|
| Claude Haiku 5.5 | 46.4% | 32.6% | estimated ± 5.9 pp, low confidence |
| SWE-1.7 | 42.3% | 28.8% | estimated ± 5.9 pp, low confidence |