Skip to content

Benchmark

CharadesSTA

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

Models measured
7
Scores recorded
9
Sources
1
IQ weight
1.97

Score distribution

59.8%
Mean
62.7%
Median
6.2%
Std dev
56%
10th percentile
63.5%
90th percentile
43.6%
Min
43.6%64.8%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.97 for chat models, measured over 7 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Qwen3 VL 235B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.64.8%
2Qwen3 VL 235Bthinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.63.5%
3qwen3-vl-30b· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.63.5%
4Qwen3 VL 32B Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.62.8%
5qwen3-vl-30bthinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.62.7%
6Qwen3 VL 32B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.61.2%
7Qwen: Qwen3 VL 8B Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.59.9%
8Qwen3 VL 8B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.56%
9Qwen2.5-VL 7B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim43.6%

Data credits: Data by LLM Stats