Saved in:
Bibliographic Details
Main Authors: Wei, Xuefeng, Wang, Zhixuan, Zhou, Xuan, Qu, Zhi, Li, Hongyao, Sakai, Yusuke, Kamigaito, Hidetaka, Watanabe, Taro
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2604.11632
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911715108585472
author Wei, Xuefeng
Wang, Zhixuan
Zhou, Xuan
Qu, Zhi
Li, Hongyao
Sakai, Yusuke
Kamigaito, Hidetaka
Watanabe, Taro
author_facet Wei, Xuefeng
Wang, Zhixuan
Zhou, Xuan
Qu, Zhi
Li, Hongyao
Sakai, Yusuke
Kamigaito, Hidetaka
Watanabe, Taro
contents We introduce CARTBENCH, a museum-grounded benchmark for evaluating vision-language models (VLMs) on Chinese artworks beyond short-form recognition and QA. CARTBENCH comprises four subtasks: CURATORQA for evidence-grounded recognition and reasoning, CATALOGCAPTION for structured four-section expert-style appreciation, REINTERPRET for defensible reinterpretation with expert ratings, and CONNOISSEURPAIRS for diagnostic authenticity discrimination under visually similar confounds. CARTBENCH is built by aligning image-bearing Palace Museum objects from Wikidata with authoritative catalog pages, spanning five art categories across multiple dynasties. Across nine representative VLMs, we find that high overall CURATORQA accuracy can mask sharp drops on hard evidence linking and style-to-period inference; long-form appreciation remains far from expert references; and authenticity-oriented diagnostic discrimination stays near chance, underscoring the difficulty of connoisseur-level reasoning for current models.
format Preprint
id arxiv_https___arxiv_org_abs_2604_11632
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle CArtBench: Evaluating Vision-Language Models on Chinese Art Understanding, Interpretation, and Authenticity
Wei, Xuefeng
Wang, Zhixuan
Zhou, Xuan
Qu, Zhi
Li, Hongyao
Sakai, Yusuke
Kamigaito, Hidetaka
Watanabe, Taro
Computation and Language
We introduce CARTBENCH, a museum-grounded benchmark for evaluating vision-language models (VLMs) on Chinese artworks beyond short-form recognition and QA. CARTBENCH comprises four subtasks: CURATORQA for evidence-grounded recognition and reasoning, CATALOGCAPTION for structured four-section expert-style appreciation, REINTERPRET for defensible reinterpretation with expert ratings, and CONNOISSEURPAIRS for diagnostic authenticity discrimination under visually similar confounds. CARTBENCH is built by aligning image-bearing Palace Museum objects from Wikidata with authoritative catalog pages, spanning five art categories across multiple dynasties. Across nine representative VLMs, we find that high overall CURATORQA accuracy can mask sharp drops on hard evidence linking and style-to-period inference; long-form appreciation remains far from expert references; and authenticity-oriented diagnostic discrimination stays near chance, underscoring the difficulty of connoisseur-level reasoning for current models.
title CArtBench: Evaluating Vision-Language Models on Chinese Art Understanding, Interpretation, and Authenticity
topic Computation and Language
url https://arxiv.org/abs/2604.11632