WebMMU: A Benchmark for Multimodal Multilingual Website Understanding and Code Generation

TL;DR

WebMMU introduces a comprehensive multilingual benchmark for web understanding and code generation, highlighting current models' strengths in information extraction but their weaknesses in reasoning, grounding, and complex code tasks.

Contribution

It unifies multiple web tasks into a single benchmark using real-world data, enabling evaluation of complex reasoning and multilingual capabilities of multimodal models.

Findings

MLLMs excel at basic info extraction.

Models struggle with reasoning and grounding.

Code editing and generation remain challenging.

Abstract

We present WebMMU, a multilingual benchmark that evaluates three core web tasks: (1) website visual question answering, (2) code editing involving HTML/CSS/JavaScript, and (3) mockup-to-code generation. Unlike prior benchmarks that treat these tasks separately, WebMMU unifies them using expert-annotated, real-world web data to assess models' abilities in complex multi-step reasoning, precise element grounding, and functional UI comprehension and coding. Our evaluation shows that while multimodal large language models (MLLMs) perform well on basic information extraction, they struggle with reasoning and grounding, editing code to preserve functionality, and generating design-to-code that maintains hierarchy and supports multilingual content. These findings reveal key limitations in current MLLMs and underscore the need for improved multimodal and cross-lingual reasoning to build future web agents capable of automating diverse web development tasks.