Calibration

AA Tau3 Banking → AA ITBench

Name: benchgap: LLM benchmark scores, measured and estimated
Creator: benchgap

AA ITBench is estimated from AA Tau3 Banking with a offset logistic curve fitted on 7 models measured on both: y = 0.5014 + (0.0000 − 0.5014) / (1 + exp(−200.00·(x − 0.5104))), R² = 0.75, cross-validated error 5.7 pp. It is used for 2 estimates.

Estimated model	AA Tau3 Banking	AA ITBench	Source
Grok 4.6	50.7%	33.3%	estimated ± 5.7 pp, low confidence
Qwen3.8-27B	48.0%	50.0%	estimated ± 5.7 pp, low confidence