CArtBench : Evaluating Vision-Language Models on Chinese Art Understanding, Interpretation, and Authenticity

Xuefeng Wei, Wang Zhixuan, Xuan Zhou, Zhi Qu, Hongyao Li, Yusuke Sakai Hidetaka Kamigaito Taro Watanabe · 2026

Read CArtBench : Evaluating Vision-Language Models on Chinese Art Understanding, Interpretation, and Authenticity by Xuefeng Wei, Wang Zhixuan, Xuan Zhou, Zhi Qu, Hongyao Li, Yusuke Sakai Hidetaka Kamigaito Taro Watanabe (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-11632.txt.

Description

Wei and seven colleagues introduce CArtBench, a museum-grounded benchmark for vision-language models on Chinese artworks beyond short QA. Built from Palace Museum objects aligned with catalog pages across five categories and multiple dynasties, it holds four subtasks. They cover evidence-grounded recognition, expert-style appreciation, reinterpretation with expert ratings, and authenticity tests under similar-looking confounds. Across nine models, high accuracy masks sharp drops on hard evidence linking and style-to-period inference. Long-form appreciation trails experts, and authenticity discrimination stays near chance.

Catalog record

CREATORXuefeng Wei, Wang Zhixuan, Xuan Zhou, Zhi Qu, Hongyao Li, Yusuke Sakai Hidetaka Kamigaito Taro Watanabe
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE64826
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.11632

Browse the full library · About the Tower of Babel