MMMU-Pro: A More Robust Multi-discipline Multimodal Understanding Benchmark

TL;DR

MMMU-Pro is a new benchmark that rigorously tests multimodal AI models' ability to understand and reason with combined visual and textual information, revealing significant performance gaps and guiding future research.

Contribution

It introduces MMMU-Pro, a more challenging and realistic multimodal understanding benchmark that filters answerable questions, adds candidate options, and incorporates vision-only inputs.

Findings

Model performance drops significantly on MMMU-Pro compared to MMMU.

Chain of Thought reasoning improves model performance.

OCR prompts have minimal impact on results.

Abstract

This paper introduces MMMU-Pro, a robust version of the Massive Multi-discipline Multimodal Understanding and Reasoning (MMMU) benchmark. MMMU-Pro rigorously assesses multimodal models' true understanding and reasoning capabilities through a three-step process based on MMMU: (1) filtering out questions answerable by text-only models, (2) augmenting candidate options, and (3) introducing a vision-only input setting where questions are embedded within images. This setting challenges AI to truly "see" and "read" simultaneously, testing a fundamental human cognitive skill of seamlessly integrating visual and textual information. Results show that model performance is substantially lower on MMMU-Pro than on MMMU, ranging from 16.8% to 26.9% across models. We explore the impact of OCR prompts and Chain of Thought (CoT) reasoning, finding that OCR prompts have minimal effect while CoT generally improves performance. MMMU-Pro provides a more rigorous evaluation tool, closely mimicking real-world scenarios and offering valuable directions for future research in multimodal AI.