Calibration
BrowseComp → GDP.pdf
GDP.pdf is estimated from BrowseComp with a offset logistic curve fitted on 6 models measured on both: y = 0.0929 + (1.2000 − 0.0929) / (1 + exp(−58.17·(x − 0.9422))), R² = 0.88, cross-validated error 6.6 pp. It is used for 14 estimates.
| Estimated model | BrowseComp | GDP.pdf | Source |
|---|---|---|---|
| Agents-A1 | 75.5% | 9.3% | estimated ± 6.6 pp, low confidence |
| Agents-A1-4B | 66.8% | 9.3% | estimated ± 6.6 pp, low confidence |
| Claude Mythos 5 | 88.0% | 12.2% | estimated ± 6.6 pp, low confidence |
| dots3-note Preview | 83.3% | 9.5% | estimated ± 6.6 pp, low confidence |
| GPT-5.4 Pro | 89.3% | 15.3% | estimated ± 6.6 pp, low confidence |
| GPT-5.5 Pro | 90.1% | 18.5% | estimated ± 6.6 pp, low confidence |
| LongCat-Flash-Lite-Sparse | 48.6% | 9.3% | estimated ± 6.6 pp, low confidence |
| Ornith-1.5-35B-A3B | 67.6% | 9.3% | estimated ± 6.6 pp, low confidence |
| Ornith-1.5-397B | 86.6% | 10.6% | estimated ± 6.6 pp, low confidence |
| Ornith-1.5-9B | 56.4% | 9.3% | estimated ± 6.6 pp, low confidence |
| Qwen3.5-27B | 61.0% | 9.3% | estimated ± 6.6 pp, low confidence |
| Qwen3.5-35B-A3B | 61.0% | 9.3% | estimated ± 6.6 pp, low confidence |
| Qwen3.5 397B | 62.0% | 9.3% | estimated ± 6.6 pp, low confidence |
| Beam | 77.4% | 9.3% | estimated ± 6.6 pp, low confidence |