benchgap
Calibration

Vals MMLU-Pro → AA-Omniscience Accuracy

AA-Omniscience Accuracy is estimated from Vals MMLU-Pro with a offset logistic curve fitted on 46 models measured on both: y = 0.2168 + (0.7140 − 0.2168) / (1 + exp(−45.99·(x − 0.8870))), R² = 0.69, cross-validated error 7.7 pp. It is used for 8 estimates.

Estimated modelVals MMLU-ProAA-Omniscience AccuracySource
Claude Haiku 4.578.7%22.2%estimated ± 7.7 pp, medium confidence
Gemini 3.1 Flash-Lite86.2%33.7%estimated ± 7.7 pp, medium confidence
GLM-4.581.2%23.2%estimated ± 7.7 pp, medium confidence
Laguna M.168.8%21.7%estimated ± 7.7 pp, low confidence
Laguna XS.269.1%21.7%estimated ± 7.7 pp, low confidence
MiMo-V2.582.9%24.9%estimated ± 7.7 pp, medium confidence
Qwen3.5 Flash84.1%27.0%estimated ± 7.7 pp, medium confidence
Qwen3.8 Max88.6%46.0%estimated ± 7.7 pp, medium confidence