Visual Aesthetic Benchmark: Can Frontier Models Judge Beauty?

Yichen Feng, Yuetai Li, Chunjiang Liu, Yuanyuan Chen, Fengqing Jiang, Yue Huang, Hang Hua, Zhengqing Yuan, Kaiyuan Zheng, Luyao Niu, Bhaskar Ramasubramanian, Basel Alomair, Xiangliang Zhang, Misha Sra, Zichen Chen, Radha Poovendran, Zhangchen Xu

arXiv:2605.12684 · 2026-05-14 공개 · arXiv · PDF

fine-tuning model-evaluation reward-models multimodal-llms visual-quality expert-annotation visual-aesthetics aesthetic-judgment

Abstract

Multimodal large language models (MLLMs) are now routinely deployed for visual understanding, generation, and curation. A substantial fraction of these applications require an explicit aesthetic judgment. Most existing solutions reduce this judgment to predicting a scalar score for a single image. We first ask whether such scores faithfully capture comparative preference: in a controlled study with eight expert annotators, score-derived rankings align poorly with the same annotators' direct comparisons, while direct ranking yields substantially higher inter-annotator agreement on best- and worst-image labels. Motivated by this finding, we introduce the Visual Aesthetic Benchmark (VAB), which casts aesthetic evaluation as comparative selection over candidate sets with matched subject matter. VAB contains 400 tasks and 1,195 images across fine art, photography, and illustration, with labels derived from the consensus of 10 independent expert judges per task. Evaluating 20 frontier MLLMs and six dedicated visual-quality reward models, we find that the strongest system identifies both the best and the worst image correctly across three random permutations of the candidate order in only 26.5% of tasks, far below the 68.9% achieved by human experts. Fine-tuning a 35B-parameter model on 2,000 expert examples brings its accuracy close to that of a 397B-parameter open-weight model, suggesting that the comparative signal in VAB is transferable. Together, these results expose a clear and measurable gap between current multimodal models and expert aesthetic judgment, and VAB provides the first set-based, expert-grounded testbed on which that gap can be tracked and closed.

한국어 요약

📋 한 줄 요약

**[멀티모달 평가 / 미적 판단]** 단일 이미지에 스칼라 점수를 매기는 기존 방식 대신, 후보 집합에서 비교 선택을 요구하는 새로운 미적 평가 벤치마크 VAB를 제안하고 최신 MLLM과 전문가 사이의 큰 격차를 정량화.

🎯 핵심 기여도

💡 핵심 아이디어

미적 평가의 본질은 절대 점수가 아니라 **비교**이며, 이 비교 신호를 매칭된 주제의 후보 집합 위에서 측정해야 모델 능력과 전문가 능력의 격차를 정직하게 드러낼 수 있다.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: 미적 판단 영역에서 현재 모델과 전문가 사이의 측정 가능한 격차를 드러내며, 그 격차를 추적·축소할 수 있는 최초의 set-based, 전문가 grounded 테스트베드를 제공. **한계**: 주제 매칭된 후보 집합에 국한된 평가이며, 자유 형식의 미적 비평·창작 가이드 같은 더 풍부한 미적 능력은 포착하지 않음.

🚀 실용적 활용