Calibration

Video-MME (w/o subtitle) → We-Math

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

We-Math is estimated from Video-MME (w/o subtitle) with a Michaelis–Menten curve fitted on 6 models measured on both: y = 2.0000·x / (1.21821 + x), R² = 0.50, cross-validated error 6.6 pp. It is used for 0 estimates.

Fitted on

Model	Video-MME (w/o subtitle)	We-Math
Claude Opus 4.5	81.4%	70.0%
Gemini 3 Pro	87.7%	86.9%
GPT-5.2	85.8%	79.0%
Kimi K2.5	83.2%	84.7%
LFM3-VL-2B	62.8%	68.6%
Qwen3.5 397B	83.7%	87.9%