Skip to content

Benchmark

VQAv2 (test)

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

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

Score distribution

75.2%
Mean
75.2%
Median
0%
Std dev
75.2%
10th percentile
75.2%
90th percentile
75.2%
Min

Leaderboard

#ModelScore
1Llama 3.2 11B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim75.2%

Data credits: Data by LLM Stats