Yuchi Wang, Rongyao Fang, Weihao Yu, Xilin Wei, Tiancheng Gu, Dingkang Yang, Hang Zhang, Dayiheng Liu, Hongsheng Li, Shuai Bai · 2026 · arXiv (Cornell University)
DOI: 10.48550/arxiv.2610.26826 · cited by 0 works (OpenAlex) · open-access PDF
에이전트의 기본 도구 사용 및 코딩 기술을 넘어, 에이전트 생성에 시각적 참조를 효과적으로 통합하는 것은 실제 생산성 작업에 AI 에이전트를 적용하는 데 똑같이 필수적입니다. 우리는 웹 페이지 개발을 통해이 중요한 기능을 연구, 기존 벤치마크가 대체로 이상화되어있는 실용적이고 쉽게 평가할 수있는 설정: 그들은 일반적으로 시각적 목표를 재구성하거나 텍스트 지침에서 웹 페이지를 생성하는 데 중점을두고, 시각적 예제에서 개발하는 더 일반적인 워크플로스를 무시합니다. 구체적으로, 우리는 RefCraft, 300 개의 현실적인 멀티모델 웹 페이지 개발 작업을 포함하는 벤치마크를 소개합니다. 그들은 자연 언어 지침을 하나에서 5 개의 참조 웹 페이지
Machine translation (Meta M2M100); the English original follows.
Beyond basic tool use and coding skills, effectively incorporating visual references into agentic generation is equally essential for applying AI agents to real-world productivity tasks. We study this crucial capability through webpage development, a practical and readily evaluable setting in which existing benchmarks remain largely idealized: they typically focus on reconstructing visual targets or generating webpages from textual instructions alone, overlooking the more common workflow of developing from visual examples. Specifically, we introduce RefCraft, a benchmark comprising 300 realistic multimodal webpage-development tasks that combine natural-language instructions with screenshots from one to five reference webpages. Agents must selectively interpret and utilize these references while rendering, visually inspecting, and iteratively refining their outputs. We evaluate the resulting open-ended artifacts using independently verifiable keypoints and complementary programmatic and VLM judges. Across nine frontier (M)LLMs, the strongest system achieves an overall score of only 52.12 out of 100, revealing that visual-reference-conditioned agentic generation remains a substantial challenge for current models in practical webpage-development settings.
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
Topics: Computer science, Benchmarking, Workflow, Focus (optics), Artificial intelligence, Web application
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.26826, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange