ChartQAPro: Results as Published

ChartQAPro asks a vision-language model to answer questions about chart images. The benchmark has 1,341 charts from 157 sources, including infographics and dashboards, and 1,948 questions, including multiple-choice, conversational, hypothetical, and unanswerable questions. Numeric answers use a relaxed match with a 5 percent margin, except years, which must match exactly. Text answers use ANLS, and multiple-choice and fact-checking answers must match exactly.

Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.