MMVU: Results as Published

MMVU tests whether a model can reason about expert-level video in four core disciplines (Science, Healthcare, Humanities and Social Sciences, and Engineering) spanning 27 subjects. The set holds 3,000 expert-annotated questions over 1,529 specialized-domain videos. Each example comes with expert-annotated reasoning rationales and domain knowledge, and each question is either open-ended or multiple-choice. The README reports accuracy; its released validation results were judged with GPT-4o, and the repository's acc_evaluation.py script produces the scores.

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.