Calibration
Agents' Last Exam → GDP.pdf
GDP.pdf is estimated from Agents' Last Exam with a Michaelis–Menten + offset curve fitted on 9 models measured on both: y = 0.0000 + 2.0000·x / (3.68756 + x), R² = 0.78, cross-validated error 3.4 pp. It is used for 7 estimates.
| Estimated model | Agents' Last Exam | GDP.pdf | Source |
|---|---|---|---|
| DeepSeek V4 Flash 0731 | 25.2% | 12.8% | estimated ± 3.4 pp, medium confidence |
| DeepSeek V4 Pro 0813 | 25.7% | 13.0% | estimated ± 3.4 pp, medium confidence |
| Gemini 3.7 Flash | 26.3% | 13.3% | estimated ± 3.4 pp, high confidence |
| Hy4 preview | 22.8% | 11.6% | estimated ± 3.4 pp, medium confidence |
| MiMo-V2.6-Flash | 27.6% | 13.9% | estimated ± 3.4 pp, high confidence |
| Qwen3.8-Flash-Next | 51.2% | 24.4% | estimated ± 3.4 pp, high confidence |
| Qwen3.8 Max | 52.4% | 24.9% | estimated ± 3.4 pp, high confidence |