multimodal benchmark
ChartMuseum
ChartMuseum is a chart question-answering benchmark of 1,162 expert-annotated questions over real-world chart images drawn from 184 sources, including academic figures, infographics, and unconventional chart designs. It specifically targets questions that require visual reasoning, such as comparing unlabeled visual elements, tracking trajectories, and judging spatial relationships.
Updated Aug 11, 2026
Models1
Model coverage1
MetricScore
EvidenceB
ChartMuseum Ranking
Higher score ranks better on this benchmark.
| 01 | AN | 86.7% | 100.0% | 1 | C |
ChartMuseum Highlights
The leading models and scores on this benchmark.
What is ChartMuseum?
What ChartMuseum measures and how its scores work.
ChartMuseum is a chart question-answering benchmark of 1,162 expert-annotated questions over real-world chart images drawn from 184 sources, including academic figures, infographics, and unconventional chart designs. It specifically targets questions that require visual reasoning, such as comparing unlabeled visual elements, tracking trajectories, and judging spatial relationships.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- ChartMuseum
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- chartmuseum|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about ChartMuseum.
Which model scores highest on ChartMuseum?
Claude Sonnet 5 is currently ranked first with 86.7%.
What does ChartMuseum measure?
ChartMuseum is a chart question-answering benchmark of 1,162 expert-annotated questions over real-world chart images drawn from 184 sources, including academic figures, infographics, and unconventional chart designs. It specifically targets questions that require visual reasoning, such as comparing unlabeled visual elements, tracking trajectories, and judging spatial relationships.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
1 model results are currently shown.
Does this benchmark affect the overall score?
No. This benchmark is shown for reference but does not contribute to the overall score.