Skip to content

Benchmark

Biomedical image analysis

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

Models measured
6
Scores recorded
8
Sources
2
IQ weight
0.68

Score distribution

60.5%
Mean
71.4%
Median
14.1%
Std dev
43.5%
10th percentile
72.2%
90th percentile
42.3%
Min
42.3%77.5%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.68 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
1GPT-6 AstraOpus 5.5 System Card · adaptive thinking · max effort unless stated· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.77.5%
2Claude Sonnet 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim72.2%
3Claude Sonnet 5.5adaptive thinking · max effort · default sampling · 5-trial mean · context ≤1M unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.72.2%
4Claude Opus 5.5Opus 5.5 System Card · adaptive thinking · max effort unless stated· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.71.4%
5Claude Opus 5Opus 5.5 System Card · adaptive thinking · max effort unless stated· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.61%
6Claude Haiku 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim43.5%
7Claude Haiku 5.5adaptive thinking · max effort · default sampling · 5-trial mean · context ≤1M unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.43.5%
8Claude Sonnet 5Opus 5.5 System Card · adaptive thinking · max effort unless stated· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.42.3%

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links. · Data by LLM Stats