dorsal/arxiv
View SchemaUR-Bench: A Benchmark for Multi-Hop Reasoning over Ultra-High-Resolution Images
| Authors | Siqi Li, Xinyu Cai, Jianbiao Mei, Nianchen Deng, Pinlong Cai, Licheng Wen, Yufan Shen, Xuemeng Yang, Botian Shi, Yong Liu |
|---|---|
| Categories | |
| ArXiv ID | 2601.08748vv1 |
| URL | https://arxiv.org/abs/2601.08748 |
| License | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ |
Abstract
Recent multimodal large language models (MLLMs) show strong capabilities in visual-language reasoning, yet their performance on ultra-high-resolution imagery remains largely unexplored. Existing visual question answering (VQA) benchmarks typically rely on medium-resolution data, offering limited visual complexity. To bridge this gap, we introduce Ultra-high-resolution Reasoning Benchmark (UR-Bench), a benchmark designed to evaluate the reasoning capabilities of MLLMs under extreme visual information. UR-Bench comprises two major categories, Humanistic Scenes and Natural Scenes, covering four subsets of ultra-high-resolution images with distinct spatial structures and data sources. Each subset contains images ranging from hundreds of megapixels to gigapixels, accompanied by questions organized into three levels, enabling evaluation of models' reasoning capabilities in ultra-high-resolution scenarios. We further propose an agent-based framework in which a language model performs reasoning by invoking external visual tools. In addition, we introduce Semantic Abstraction and Retrieval tools that enable more efficient processing of ultra-high-resolution images. We evaluate state-of-the-art models using both an end-to-end MLLMs and our agent-based framework, demonstrating the effectiveness of our framework.
{
"annotation_id": "9d1905a7-7682-4443-8ae0-a968bb28992e",
"date_created": "2026-02-17T05:53:15.890000Z",
"date_modified": "2026-02-17T05:53:15.890000Z",
"file_hash": "5757d19bd3261d90e5725ebdc41289fc0352e6f7983066767202a5c3e762797a",
"private": false,
"record": {
"abstract": "Recent multimodal large language models (MLLMs) show strong capabilities in visual-language reasoning, yet their performance on ultra-high-resolution imagery remains largely unexplored. Existing visual question answering (VQA) benchmarks typically rely on medium-resolution data, offering limited visual complexity. To bridge this gap, we introduce Ultra-high-resolution Reasoning Benchmark (UR-Bench), a benchmark designed to evaluate the reasoning capabilities of MLLMs under extreme visual information. UR-Bench comprises two major categories, Humanistic Scenes and Natural Scenes, covering four subsets of ultra-high-resolution images with distinct spatial structures and data sources. Each subset contains images ranging from hundreds of megapixels to gigapixels, accompanied by questions organized into three levels, enabling evaluation of models\u0027 reasoning capabilities in ultra-high-resolution scenarios. We further propose an agent-based framework in which a language model performs reasoning by invoking external visual tools. In addition, we introduce Semantic Abstraction and Retrieval tools that enable more efficient processing of ultra-high-resolution images. We evaluate state-of-the-art models using both an end-to-end MLLMs and our agent-based framework, demonstrating the effectiveness of our framework.",
"arxiv_id": "2601.08748",
"authors": [
"Siqi Li",
"Xinyu Cai",
"Jianbiao Mei",
"Nianchen Deng",
"Pinlong Cai",
"Licheng Wen",
"Yufan Shen",
"Xuemeng Yang",
"Botian Shi",
"Yong Liu"
],
"categories": [
"cs.CV",
"cs.AI"
],
"license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
"title": "UR-Bench: A Benchmark for Multi-Hop Reasoning over Ultra-High-Resolution Images",
"url": "https://arxiv.org/abs/2601.08748",
"version": "v1"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "b1d78c85-ab6b-4bf3-8722-e2d435e026ee",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}