Needle In A Multimodal Haystack

TL;DR

This paper introduces MM-NIAH, a benchmark for evaluating multimodal large language models' ability to understand long multimodal documents across retrieval, counting, and reasoning tasks, revealing significant room for improvement.

Contribution

The paper presents the first benchmark specifically designed to assess MLLMs' comprehension of long multimodal content, filling a critical evaluation gap.

Findings

Existing MLLMs perform poorly on long multimodal tasks.

Models show especially weak performance on vision-centric tasks.

The benchmark provides a new platform for future research.

Abstract

With the rapid advancement of multimodal large language models (MLLMs), their evaluation has become increasingly comprehensive. However, understanding long multimodal content, as a foundational ability for real-world applications, remains underexplored. In this work, we present Needle In A Multimodal Haystack (MM-NIAH), the first benchmark specifically designed to systematically evaluate the capability of existing MLLMs to comprehend long multimodal documents. Our benchmark includes three types of evaluation tasks: multimodal retrieval, counting, and reasoning. In each task, the model is required to answer the questions according to different key information scattered throughout the given multimodal document. Evaluating the leading MLLMs on MM-NIAH, we observe that existing models still have significant room for improvement on these tasks, especially on vision-centric evaluation. We hope this work can provide a platform for further research on long multimodal document comprehension and contribute to the advancement of MLLMs. Code and benchmark are released at https://github.com/OpenGVLab/MM-NIAH.