MMDeepResearch-Bench: A Benchmark for Multimodal Deep Research Agents

TL;DR

This paper introduces MMDeepResearch-Bench, a comprehensive benchmark with 140 tasks across 21 domains to evaluate multimodal deep research agents in generating citation-rich, evidence-grounded reports connecting images and text.

Contribution

It presents a new end-to-end multimodal benchmark and an interpretable evaluation pipeline for assessing report quality, citation accuracy, and visual-text integrity in deep research agents.

Findings

Models show trade-offs between report quality and evidence fidelity.

Strong prose does not ensure faithful evidence use.

Multimodal integrity is a key bottleneck for research agents.

Abstract

Deep Research Agents (DRAs) generate citation-rich reports via multi-step search and synthesis, yet existing benchmarks mainly target text-only settings or short-form multimodal QA, missing end-to-end multimodal evidence use. We introduce MMDeepResearch-Bench (MMDR-Bench), a benchmark of 140 expert-crafted tasks across 21 domains, where each task provides an image-text bundle to evaluate multimodal understanding and citation-grounded report generation. Compared to prior setups, MMDR-Bench emphasizes report-style synthesis with explicit evidence use, where models must connect visual artifacts to sourced claims and maintain consistency across narrative, citations, and visual references. We further propose a unified, interpretable evaluation pipeline: Formula-LLM Adaptive Evaluation (FLAE) for report quality, Trustworthy Retrieval-Aligned Citation Evaluation (TRACE) for citation-grounded evidence alignment, and Multimodal Support-Aligned Integrity Check (MOSAIC) for text-visual integrity, each producing fine-grained signals that support error diagnosis beyond a single overall score. Experiments across 25 state-of-the-art models reveal systematic trade-offs between generation quality, citation discipline, and multimodal grounding, highlighting that strong prose alone does not guarantee faithful evidence use and that multimodal integrity remains a key bottleneck for deep research agents.