Calibration

RealWorldQA → MMMU-Pro

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

MMMU-Pro is estimated from RealWorldQA with a Hill curve fitted on 5 models measured on both: y = 0.0000 + (1.2000 − 0.0000)·x^6.00 / (0.79332^6.00 + x^6.00), R² = 0.84, cross-validated error 11.4 pp. It is used for 2 estimates.

Estimated model	RealWorldQA	MMMU-Pro	Source
North Micro Vision Instruct	62.2%	22.6%	estimated ± 11.4 pp, low confidence
Ternary Bonsai 2 27B	80.1%	61.8%	estimated ± 11.4 pp, low confidence