@inproceedings{zhao-etal-2026-vtcbench,
title = "{VTCB}ench: Can Vision-Language Models Understand Long Contexts with Vision-Text Compression?",
author = "Zhao, Hongbo and
Wang, Meng and
Zhu, Fei and
Liu, Wenzhuo and
Ni, Bolin and
Zeng, Fanhu and
Meng, Gaofeng and
Zhang, Zhaoxiang and
Liu, Cheng-Lin",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.145/",
pages = "3349--3379",
ISBN = "979-8-89176-459-0",
abstract = "The computational and memory overheads associated with expanding the context window of LLMs severely limit their scalability. A noteworthy solution is vision-text compression (VTC), exemplified by frameworks like DeepSeek-OCR and Glyph, which convert long texts into dense 2D visual representations, thereby achieving token compression ratios of 2x{--}10x. However, the impact of this high information density on the core long-context capabilities of vision-language models (VLMs) remains under-investigated. To address this gap, we introduce the first benchmark for VTC and systematically assess the performance of VLMs across three long-context understanding settings: VTC-Retrieval, VTC-Reasoning, and VTC-Memory. We further introduce VTCBench-Wild to evaluate robustness across diverse rendering configurations. We comprehensively evaluate leading open-weight and proprietary models on our benchmarks. The results indicate that several VLMs can retrieve information from rendered text in short-to-moderate contexts, but this ability degrades at longer lengths, and the degradation is especially pronounced on tasks requiring binding and integrating dispersed evidence. This study provides diagnostic evidence of VTC and serves as a foundation for designing more efficient and scalable VLMs."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhao-etal-2026-vtcbench">
<titleInfo>
<title>VTCBench: Can Vision-Language Models Understand Long Contexts with Vision-Text Compression?</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hongbo</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Meng</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fei</namePart>
<namePart type="family">Zhu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wenzhuo</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bolin</namePart>
<namePart type="family">Ni</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fanhu</namePart>
<namePart type="family">Zeng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gaofeng</namePart>
<namePart type="family">Meng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhaoxiang</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Cheng-Lin</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>The computational and memory overheads associated with expanding the context window of LLMs severely limit their scalability. A noteworthy solution is vision-text compression (VTC), exemplified by frameworks like DeepSeek-OCR and Glyph, which convert long texts into dense 2D visual representations, thereby achieving token compression ratios of 2x–10x. However, the impact of this high information density on the core long-context capabilities of vision-language models (VLMs) remains under-investigated. To address this gap, we introduce the first benchmark for VTC and systematically assess the performance of VLMs across three long-context understanding settings: VTC-Retrieval, VTC-Reasoning, and VTC-Memory. We further introduce VTCBench-Wild to evaluate robustness across diverse rendering configurations. We comprehensively evaluate leading open-weight and proprietary models on our benchmarks. The results indicate that several VLMs can retrieve information from rendered text in short-to-moderate contexts, but this ability degrades at longer lengths, and the degradation is especially pronounced on tasks requiring binding and integrating dispersed evidence. This study provides diagnostic evidence of VTC and serves as a foundation for designing more efficient and scalable VLMs.</abstract>
<identifier type="citekey">zhao-etal-2026-vtcbench</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.145/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>3349</start>
<end>3379</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T VTCBench: Can Vision-Language Models Understand Long Contexts with Vision-Text Compression?
%A Zhao, Hongbo
%A Wang, Meng
%A Zhu, Fei
%A Liu, Wenzhuo
%A Ni, Bolin
%A Zeng, Fanhu
%A Meng, Gaofeng
%A Zhang, Zhaoxiang
%A Liu, Cheng-Lin
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F zhao-etal-2026-vtcbench
%X The computational and memory overheads associated with expanding the context window of LLMs severely limit their scalability. A noteworthy solution is vision-text compression (VTC), exemplified by frameworks like DeepSeek-OCR and Glyph, which convert long texts into dense 2D visual representations, thereby achieving token compression ratios of 2x–10x. However, the impact of this high information density on the core long-context capabilities of vision-language models (VLMs) remains under-investigated. To address this gap, we introduce the first benchmark for VTC and systematically assess the performance of VLMs across three long-context understanding settings: VTC-Retrieval, VTC-Reasoning, and VTC-Memory. We further introduce VTCBench-Wild to evaluate robustness across diverse rendering configurations. We comprehensively evaluate leading open-weight and proprietary models on our benchmarks. The results indicate that several VLMs can retrieve information from rendered text in short-to-moderate contexts, but this ability degrades at longer lengths, and the degradation is especially pronounced on tasks requiring binding and integrating dispersed evidence. This study provides diagnostic evidence of VTC and serves as a foundation for designing more efficient and scalable VLMs.
%U https://aclanthology.org/2026.emnlp-main.145/
%P 3349-3379
Markdown (Informal)
[VTCBench: Can Vision-Language Models Understand Long Contexts with Vision-Text Compression?](https://aclanthology.org/2026.emnlp-main.145/) (Zhao et al., EMNLP 2026)
ACL
- Hongbo Zhao, Meng Wang, Fei Zhu, Wenzhuo Liu, Bolin Ni, Fanhu Zeng, Gaofeng Meng, Zhaoxiang Zhang, and Cheng-Lin Liu. 2026. VTCBench: Can Vision-Language Models Understand Long Contexts with Vision-Text Compression?. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 3349–3379, Budapest, Hungary. Association for Computational Linguistics.