/

← All benchmarks

AI2D

vision · diagram-understanding↗ Quelle

44Evidence

active

What it measures

Verstaendnis von Lehrbuchdiagrammen und deren Beschriftungen.

What it does not measure

Fotos, Video.

How the score is composed

Currency82

Discrimination29

Headroom62

Contamination safety15

Gaming resistance55

Methodological rigor50

Reproducibility60

Details

Benchmaxxing risk: medium
Difficulty: low
Public test set: yes
Human-verified: no
LLM-judged: no
Letzter Messwert: 2026-07-25
Skala: 0–100 % correct (höher besser)

Best measured models

1Mage-VLMicrosoftvendor number81.54 % correct
2Phi-4-multimodalMicrosoftvendor number78.1 % correct