Vision2Web: A Hierarchical Benchmark for Visual Website Development with Agent Verification

TL;DR

Vision2Web is a comprehensive benchmark for evaluating visual website development capabilities of AI models, covering static, interactive, and full-stack tasks, with a new verification paradigm.

Contribution

It introduces a hierarchical benchmark with real-world tasks and a novel agent verification method for systematic evaluation of website development models.

Findings

State-of-the-art models show significant performance gaps.

Models struggle with full-stack website development.

Benchmark covers 193 tasks across 16 categories.

Abstract

Recent advances in large language models have improved the capabilities of coding agents, yet systematic evaluation of complex, end-to-end website development remains limited. To address this gap, we introduce Vision2Web, a hierarchical benchmark for visual website development, spanning from static UI-to-code generation, interactive multi-page frontend reproduction, to long-horizon full-stack website development. The benchmark is constructed from real-world websites and comprises a total of 193 tasks across 16 categories, with 918 prototype images and 1,255 test cases. To support flexible, thorough and reliable evaluation, we propose workflow-based agent verification paradigm based on two complementary components: a GUI agent verifier and a VLM-based judge. We evaluate multiple visual language models instantiated under different coding-agent frameworks, revealing substantial performance gaps at all task levels, with state-of-the-art models still struggling on full-stack development.