Calibration
GPQA → Vals GPQA Diamond
Vals GPQA Diamond is estimated from GPQA with a inverse Michaelis–Menten curve fitted on 25 models measured on both: y = 1.39188·(x − 0.1281) / (0.1281 + 2.0000 − x), R² = 0.62, cross-validated error 4.8 pp. It is used for 25 estimates.
| Estimated model | GPQA | Vals GPQA Diamond | Source |
|---|---|---|---|
| Claude 3.5 Sonnet | 59.4% | 42.3% | estimated ± 4.8 pp, medium confidence |
| Claude Opus 4.5 | 87.0% | 82.1% | estimated ± 4.8 pp, high confidence |
| DeepSeek V3 | 59.1% | 41.9% | estimated ± 4.8 pp, medium confidence |
| Gemini 2.5 Pro | 83.0% | 75.3% | estimated ± 4.8 pp, high confidence |
| Gemma 4 E2B | 43.4% | 25.1% | estimated ± 4.8 pp, medium confidence |
| Gemma 4 E4B | 58.6% | 41.3% | estimated ± 4.8 pp, medium confidence |
| GPT-4.1 | 66.3% | 50.8% | estimated ± 4.8 pp, medium confidence |
| GPT-4.1 mini | 64.2% | 48.1% | estimated ± 4.8 pp, medium confidence |
| GPT-4.1 nano | 50.3% | 32.1% | estimated ± 4.8 pp, medium confidence |
| Granite 4.2 30B | 66.4% | 51.0% | estimated ± 4.8 pp, medium confidence |
| Granite 4.2 3B | 54.8% | 37.0% | estimated ± 4.8 pp, medium confidence |
| Granite 4.2 8B | 64.1% | 48.1% | estimated ± 4.8 pp, medium confidence |
| Kimi K2.5 (Reasoning) | 87.6% | 83.1% | estimated ± 4.8 pp, high confidence |
| LFM2.5-VL-450M | 25.7% | 9.6% | estimated ± 4.8 pp, medium confidence |
| Ling 2.6 Flash | 59.0% | 41.8% | estimated ± 4.8 pp, medium confidence |
| o1 | 75.7% | 63.8% | estimated ± 4.8 pp, medium confidence |
| o1-pro | 79.0% | 68.8% | estimated ± 4.8 pp, medium confidence |
| o3-mini | 77.2% | 66.1% | estimated ± 4.8 pp, medium confidence |
| Qwen3 235B 2507 | 77.5% | 66.5% | estimated ± 4.8 pp, medium confidence |
| Qwen3.5-122B-A10B | 86.6% | 81.4% | estimated ± 4.8 pp, high confidence |
| Qwen3.5-27B | 85.5% | 79.5% | estimated ± 4.8 pp, high confidence |
| Qwen3.5-35B-A3B | 84.2% | 77.3% | estimated ± 4.8 pp, high confidence |
| Qwen3.5 397B | 88.4% | 84.6% | estimated ± 4.8 pp, high confidence |
| Qwen3.6-27B | 87.8% | 83.5% | estimated ± 4.8 pp, high confidence |
| Qwen3.6-35B-A3B | 86.0% | 80.3% | estimated ± 4.8 pp, high confidence |