Alibaba's Qwen3.8-Max Secures Second Place in Image-to-WebDev Arena
Alibaba's newly released Qwen3.8-Max model has achieved the second-place position on the Image-to-WebDev Arena leaderboard. Scoring 1,631 points, it trails the top-ranked Claude Opus 5 (Max) by only 39 points. This achievement highlights Qwen3.8-Max's strong capabilities in multimodal vision-to-code tasks, showing that open-weight models are closing the gap with proprietary giants. It offers developers a highly competitive alternative for automated web design and front-end code generation. Qwen3.8-Max is a massive 2.4 trillion parameter Mixture of Experts (MoE) model featuring a 1-million-token context window. The Image-to-WebDev benchmark evaluates models by having them generate functional web applications from UI screenshots, wireframes, or design mockups.
## BACKGROUND
The Image-to-WebDev Arena is a specialized leaderboard where users upload design images and vote on which of two anonymous AI models generates the best working clone. Qwen is a family of large language models developed by Alibaba Cloud, known for releasing highly capable open-weight models that compete with proprietary systems.