arXiv cs.CL· Haoyu Sun, Huichen Will Wang, Jiawei Gu, Linjie Li, Yu Cheng·· 13 小时前AI 评分35
FullFront：覆盖完整前端工程流程的 MLLM 基准测试
FullFront: Benchmarking MLLMs Across the Full Front-End Engineering Workflow
AI 导读
FullFront 是一个评估多模态大模型（MLLM）完整前端开发流程的基准，涵盖网页设计、网页感知问答和网页代码生成三项任务。它通过两阶段流程将真实网页转换为干净、标准化的 HTML，避免代码臃肿与版权问题。测试显示当前 SOTA MLLM 在页面感知、代码生成（尤其是图像处理与布局）及交互实现上存在明显局限，与人类专家水平差距显著。
正文 · 原文
Abstract:Front-end engineering involves a complex workflow where engineers conceptualize designs, translate them into code, and iteratively refine the implementation. While recent benchmarks primarily focus on converting visual designs to code, we present FullFront, a benchmark designed to evaluate Multimodal Large Language Models (MLLMs) \textbf{across the full front-end development pipeline}. FullFront assesses three fundamental tasks that map directly to the front-end engineering pipeline: Webpage Design (conceptualization phase), Webpage Perception QA (comprehension of visual organization and elements), and Webpage Code Generation (implementation phase). Unlike existing benchmarks that use either scraped websites with bloated code or oversimplified LLM-generated HTML, FullFront employs a novel, two-stage process to transform real-world webpages into clean, standardized HTML while maintaining diverse visual designs and avoiding copyright issues. Extensive testing of state-of-the-art MLLMs reveals significant limitations in page perception, code generation (particularly for image handling and layout), and interaction implementation. Our results quantitatively demonstrate performance disparities across models and tasks, and highlight a substantial gap between current MLLM capabilities and human expert performance in front-end engineering. The FullFront benchmark and code are available in this https URL.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2505.17399 [cs.CL] |
| (or arXiv:2505.17399v4 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2505.17399 arXiv-issued DOI via DataCite |
Submission history
From: Haoyu Sun [view email]
[v1]
Fri, 23 May 2025 02:16:11 UTC (7,701 KB)
[v2]
Mon, 26 May 2025 11:15:36 UTC (7,701 KB)
[v3]
Mon, 5 Oct 2026 09:16:43 UTC (9,735 KB)
[v4]
Tue, 27 Oct 2026 06:32:22 UTC (9,734 KB)
来源：arXiv cs.CL · arxiv.org