Calibration

Video-MME (w/o subtitle) → ERQA

Name: benchgap: LLM benchmark scores, measured and estimated
Creator: benchgap
License: https://creativecommons.org/licenses/by-nc/4.0/

ERQA is estimated from Video-MME (w/o subtitle) with a Michaelis–Menten curve fitted on 5 models measured on both: y = 2.0000·x / (1.92798 + x), R² = 0.42, cross-validated error 9.3 pp. It is used for 0 estimates.

Fitted on

Model	Video-MME (w/o subtitle)	ERQA
Claude Opus 4.5	81.4%	46.8%
Gemini 3 Pro	87.7%	70.5%
GPT-5.2	85.8%	59.8%
LFM3-VL-2B	62.8%	48.2%
Qwen3.5 397B	83.7%	67.5%