Skip to content

Benchmark

BLINK - Jigsaw

Vision benchmarks measure how well models understand images. Scores are percentages.

Models measured
4
Scores recorded
4
Sources
1
IQ weight
No weight published

Score distribution

36.7%
Mean
54%
Median
21.4%
Std dev
3.3%
10th percentile
56.7%
90th percentile
3.3%
Min

Leaderboard

#ModelScore
1LLaVA 1.6 34BJigsaw· EvalAI data56.7%
2Gemini ProJigsaw· EvalAI data54%
3Claude 3 OpusJigsaw· EvalAI data32.7%
4Qwen-VL MaxJigsaw· EvalAI data3.3%