Skip to content

Benchmark

InfoVQA

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

Models measured
6
Scores recorded
6
Sources
1
IQ weight
0.47

Score distribution

69.8%
Mean
72.7%
Median
10.5%
Std dev
64.9%
10th percentile
82.6%
90th percentile
50%
Min
50%82.6%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.47 for chat models, measured over 6 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Qwen2.5-VL 7B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim82.6%
2DeepSeek VL2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim78.1%
3Phi 4 Multimodal Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim72.7%
4Gemma 3 27B IT· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim70.6%
5Gemma 3 12B IT· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim64.9%
6Gemma 3 4B IT· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim50%

Data credits: Data by LLM Stats