analiza aiqbench
Vision2Web introduces a rigorous hierarchical benchmark to evaluate coding agents on complex, end-to-end website development tasks ranging from static UI generation to full-stack implementation. By employing a workflow-based verification paradigm combining GUI agents and visual language models, the study exposes significant performance gaps where current state-of-the-art systems still struggle with long-horizon projects. This work provides essential infrastructure for advancing autonomous coding capabilities within AI research communities focused on software engineering applications.
Vision2Web: A Hierarchical Benchmark for Visual Website Development with Agent Verification
arXiv:2603.26648v3 Announce Type: replace-cross 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.
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.
Software Engineering (cs.SE); Artificial Intelligence (cs.AI)
arXiv:2603.26648 [cs.SE]
(or arXiv:2603.26648v3 [cs.SE] for this version)
Text integral sub licența deschisă a publicației · 143 cuvinte
Subiecte conexe
[AINews] Quasi-Riemann-Hypothesis: OpenAI publishes 722 math papers solving 90 of the top 500 open math problems; “the most significant moment” in >100 years of mathematics
Latent Space · 7 oct. 2026
ViCo: Visual-oriented Coding with Self-Reflection for Chart Replication
arXiv · 16 sept. 2026
AgentPProf: Semantic Profiler for Long Horizon AI Agents
arXiv · 18 sept. 2026
[AINews] not much happened today
Latent Space · 18 sept. 2026