The Year in AI Papers
The year/Independent research

Paper 2602.02185

Vision-DeepResearch Benchmark: Rethinking Visual and Textual Search for Multimodal Large Language Models

Yu Zeng, Wenxuan Huang, Zhen Fang, Shuang Chen, Yufan Shen, Yishuo Cai, Xiaoman Wang, Zhenfei Yin, Lin Chen, Zehui Chen, Shiting Huang, Yiming Zhao, et al.

Published
Feb 2026
Research lab
Independent
Citations
21
GitHub
Not linked

01 In brief

Summary

The paper introduces VDR-Bench, a benchmark of 2,000 VQA instances designed to evaluate Vision-DeepResearch systems under realistic conditions.

It identifies two flaws in existing benchmarks: they are not visual-search-centric (answers can be inferred from text cues or model priors) and rely on idealized retrieval (near-exact whole-image matching).

VDR-Bench is built via a multi-stage pipeline involving manual cropping, visual entity verification, knowledge-graph-based multi-hop question expansion, and solvability checks.

Experiments show that models perform poorly without search (e.g., Gemini 2.5 Pro: 8.2% accuracy), but improve with cropped-image search and text search (CIS+TS), and further with Multi-turn Visual Forcing (MVF), which encourages iterative cropping and cross-modal verification.

The best overall accuracy was 30.0% for Gemini 2.5 Pro with MVF.

The paper also proposes a multi-round cropped-search workflow that significantly boosts performance, highlighting the importance of iterative visual search and entity-level retrieval for multimodal deep research.

02 From the paper

Abstract

Multimodal Large Language Models (MLLMs) have advanced VQA and now support Vision-DeepResearch systems that use search engines for complex visual-textual fact-finding. However, evaluating these visual and textual search abilities is still difficult, and existing benchmarks have two major limitations. First, existing benchmarks are not visual search-centric: answers that should require visual search are often leaked through cross-textual cues in the text questions or can be inferred from the prior world knowledge in current MLLMs. Second, overly idealized evaluation scenario: On the image-search side, the required information can often be obtained via near-exact matching against the full image, while the text-search side is overly direct and insufficiently challenging. To address these issues, we construct the Vision-DeepResearch benchmark (VDR-Bench) comprising 2,000 VQA instances. All questions are created via a careful, multi-stage curation pipeline and rigorous expert review, designed to assess the behavior of Vision-DeepResearch systems under realistic real-world conditions. Moreover, to address the insufficient visual retrieval capabilities of current MLLMs, we propose a simple multi-round cropped-search workflow. This strategy is shown to effectively improve model performance in realistic visual retrieval scenarios. Overall, our results provide practical guidance for the design of future multimodal deep-research systems. The code will be released in https://github.com/Osilly/Vision-DeepResearch.