Calibration
MMMU-Pro → GDP.pdf
GDP.pdf is estimated from MMMU-Pro with a Hill curve fitted on 17 models measured on both: y = 0.0000 + (1.2000 − 0.0000)·x^6.00 / (1.05153^6.00 + x^6.00), R² = 0.74, cross-validated error 4.0 pp. It is used for 3 estimates.
| Estimated model | MMMU-Pro | GDP.pdf | Source |
|---|---|---|---|
| Claude Opus 5.5 (xhigh with fallback) | 87.0% | 29.1% | estimated ± 4.0 pp, high confidence |
| Claude Opus 5.5 (high with fallback) | 86.0% | 27.6% | estimated ± 4.0 pp, high confidence |
| Claude Opus 5.5 (medium with fallback) | 86.0% | 27.6% | estimated ± 4.0 pp, high confidence |