eric zakariasson(@ericzakariasson)
just shipped web ui bench! measuring taste is hard, so i gave 20 models the same ui component promp...
5.2Score
TL;DR · AI 摘要
Eric Zakariasson 发布 Web UI Bench，用统一 UI 组件提示词测试 20 个模型输出，支持并排对比，但未披露评测维度、数据集或方法论细节。
核心要点
- 仅提供主观视觉对比，缺乏量化指标与基准定义
- 强调‘品味难以衡量’，但未说明如何控制 prompt 差异或渲染一致性
- 工具基于 Cursor SDK 构建，属开发者实验性产物，非标准化评测框架
思维导图
用一张图看清主题之间的关系。
查看大纲文本（无障碍 / 无 JS 友好）
- Web UI Bench
#AI#UI生成#模型评测#Cursor
打开原文measuring taste is hard, so i gave 20 models the same ui component prompts and put every output side by side so you can compare them yourself. let me know which you think is best!
built with cursor sdk
https://t.co/fYaqoSqVLg https://t.co/yz3pAExrG5" / X
eric zakariasson on X: "just shipped web ui bench! measuring taste is hard, so i gave 20 models the same ui component prompts and put every output side by side so you can compare them yourself. let me know which you think is best! built with cursor sdk https://t.co/fYaqoSqVLg https://t.co/yz3pAExrG5" / X
Don’t miss what’s happening
eric zakariasson 
just shipped web ui bench! measuring taste is hard, so i gave 20 models the same ui component prompts and put every output side by side so you can compare them yourself. let me know which you think is best! built with cursor sdk https://webuibench.dev
0:07
·
68
54
844
676
Read 68 replies