Share your thoughts, 1 month free Claude Pro on usSee more

Visual Language Model Evaluation on MM-Bench EN

65.8MM-Bench (EN) Score

Vanilla

Updated 13d ago

Evaluation Results

Method	Links
Vanilla 2026.05		65.8	100
DART 2026.05		65.4	96.4
SPpruner 2026.05		65	97.7
DivPrune 2026.05		64.2	96.1
VisionZip 2026.05		63.1	96.2
FastV 2026.05		61.6	86.8
PyramidDrop 2026.05		57.9	90.1