MMMU-Pro
The harder MMMU-Pro multimodal reasoning benchmark (college-level subject tasks with text and images); the variant current frontier models report.
What this benchmark measures
The harder MMMU-Pro multimodal reasoning benchmark (college-level subject tasks with text and images); the variant current frontier models report.
Rows on this page are sourced from public benchmark artifacts, leaderboard exports, or source-linked model reports. Each row keeps benchmark version, source model name, and available run details attached to the score.
The metric shown here is accuracy. It should be interpreted within MMMU-Pro, not compared as part of a site-wide ranking.
What to be careful about
Preserve multimodal prompt packaging and image handling as run context.
Frequently asked
What is MMMU-Pro?
The harder MMMU-Pro multimodal reasoning benchmark (college-level subject tasks with text and images); the variant current frontier models report. It is a multimodal benchmark measured by accuracy.
What does accuracy mean on MMMU-Pro?
MMMU-Pro reports accuracy (%); higher is better. Scores are shown only within MMMU-Pro and are never averaged with other benchmarks.
What is the top reported MMMU-Pro score?
GPT-5.6 Sol has the top reported score on MMMU-Pro: 83.0% (accuracy).
Why do MMMU-Pro scores differ across runs?
Harness, scaffold, reasoning effort, and prompt setup change results, so two runs of the same model can differ. evals.report keeps each score with its run context so the differences stay visible.
Does evals.report rank models across benchmarks?
No. MMMU-Pro scores are shown within their own metric; evals.report never combines benchmarks into a composite ranking or a single "best model".