Calibration

VideoMMMU → DynaMath

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

DynaMath is estimated from VideoMMMU with a linear curve fitted on 7 models measured on both: y = 0.1543·x + 0.7146, R² = 0.46, cross-validated error 5.7 pp. It is used for 0 estimates.

Fitted on

Model	VideoMMMU	DynaMath
Claude Opus 4.5	84.4%	79.7%
Gemini 3 Pro	87.6%	85.1%
GPT-5.2	85.9%	86.8%
Kimi K2.5	86.6%	84.4%
LFM3-VL-2B	49.6%	79.2%
Qwen3.5 397B	84.7%	86.3%
Qwen3.6-27B	84.4%	85.6%