Calibration
BrowseComp → GDPval-AA
GDPval-AA is estimated from BrowseComp with a Hill curve fitted on 34 models measured on both: y = 0.1798 + (1.2000 − 0.1798)·x^6.00 / (1.00344^6.00 + x^6.00), R² = 0.73, cross-validated error 7.8 pp. It is used for 10 estimates.
| Estimated model | BrowseComp | GDPval-AA | Source |
|---|---|---|---|
| GPT-5.5 Pro | 90.1% | 53.1% | estimated ± 7.8 pp, medium confidence |
| GPT-5.4 Pro | 89.3% | 51.8% | estimated ± 7.8 pp, medium confidence |
| Nemotron 3.5 Bolt 74B-A9B BF16 | 70.0% | 28.5% | estimated ± 7.8 pp, medium confidence |
| Qwen3.5 397B | 69.0% | 27.7% | estimated ± 7.8 pp, medium confidence |
| Agents-A1-4B | 66.8% | 26.2% | estimated ± 7.8 pp, medium confidence |
| Qwen3.5-27B | 61.0% | 22.9% | estimated ± 7.8 pp, medium confidence |
| Qwen3.5-35B-A3B | 61.0% | 22.9% | estimated ± 7.8 pp, medium confidence |
| Sarvam 105B | 49.5% | 19.4% | estimated ± 7.8 pp, medium confidence |
| LongCat-Flash-Lite-Sparse | 48.6% | 19.3% | estimated ± 7.8 pp, medium confidence |
| Sarvam 30B | 35.5% | 18.2% | estimated ± 7.8 pp, low confidence |
Fitted on
| Model | BrowseComp | GDPval-AA |
|---|---|---|
| Claude Opus 4.7 | 79.3% | 42.6% |
| Claude Opus 4.8 | 84.3% | 47.7% |
| Claude Opus 5 | 90.8% | 61.1% |
| Claude Sonnet 5 | 84.7% | 48.2% |
| DeepSeek V4 Flash 0731 | 73.2% | 46.7% |
| DeepSeek V4 Pro | 83.4% | 54.5% |
| GLM-4.7 | 52.0% | 25.7% |
| GLM-5.1 | 68.0% | 30.8% |
| GPT-5.4 | 82.7% | 37.3% |
| GPT-5.5 | 84.4% | 42.5% |
| GPT-5.6 Luna | 83.3% | 48.0% |
| GPT-5.6 Sol | 92.2% | 55.4% |
| GPT-5.6 Terra | 87.5% | 47.6% |
| GPT-6 Astra | 91.5% | 53.7% |
| Inkling | 77.1% | 28.6% |
| Inkling-Small | 77.4% | 31.2% |
| Kimi K2.6 | 83.2% | 26.9% |
| Kimi K3.1 Pro | 92.1% | 67.3% |
| Kimi K2.5 | 60.6% | 17.0% |
| Kimi K3 | 91.2% | 51.6% |
| Ling 3.0 Flash | 72.2% | 22.4% |
| Ling 3.1 Flash | 91.7% | 55.9% |
| MiniMax M3 | 83.5% | 37.1% |
| Mistral Large 4 | 84.5% | 46.1% |
| Nemotron 3.5 Bolt 74B-A9B | 66.8% | 24.0% |
| Nemotron 3.5 Lightning 30B A3B NVFP4 | 36.8% | 6.9% |
| Nemotron 3.5 Super VL 120B A12B | 78.0% | 33.1% |
| Nemotron 3.5 Super VL 120B A12B NVFP4 | 77.5% | 33.1% |
| Nemotron 3 Ultra | 44.4% | 33.1% |
| Qwen3.5-122B-A10B | 63.8% | 15.7% |
| Beam 1.0 501B-A23B | 76.8% | 33.9% |
| Solar Pro 4 | 49.2% | 30.3% |
| Step 3.7 Flash | 75.8% | 25.8% |
| Step 5 Preview | 88.7% | 54.3% |