benchgap
Calibration

SimpleVQA → MMMU-Pro

MMMU-Pro is estimated from SimpleVQA with a Hill curve fitted on 16 models measured on both: y = 0.0000 + (0.8130 − 0.0000)·x^6.00 / (0.36526^6.00 + x^6.00), R² = 0.94, cross-validated error 5.1 pp. It is used for 0 estimates.

Fitted on

ModelSimpleVQAMMMU-Pro
Claude Opus 4.565.7%70.6%
dots3-note Preview72.5%79.1%
Gemini 3.1 Pro72.4%83.9%
Gemini 3 Pro73.2%81.0%
GPT-5.255.8%79.5%
GPT-5.461.1%81.2%
Grok 4.2057.4%75.2%
Kimi K2.571.2%78.5%
LFM2.5-VL-3B35.4%30.5%
LFM3-VL-2B36.3%45.7%
Muse Spark71.3%80.4%
Qwen3.5 397B67.1%79.0%
Qwen3.6-27B56.1%75.8%
Qwen3.6-35B-A3B58.9%75.3%
Qwen3.7 Plus81.7%79.0%
Qwen3.8 Max75.0%82.3%