Exploring AI Performance in UI/UX and Coding: Insights from a Global Survey

In recent months, I embarked on a comprehensive research project to evaluate how various AI models perform in designing user interfaces, enhancing user experience, and generating code. Drawing from a global survey involving thousands of users, I aimed to identify which models stand out and where they still fall short.

A Transparent and Open Approach

All data collected, along with the AI-generated outputs, are sourced from open-source models and free-to-use tools. My intention is purely to share insights and research findings—no financial gain involved.

Introducing the Crowd-Sourced Benchmark Platform

To facilitate this analysis, I developed a platform called Design Arena. This community-driven tool allows users to generate websites, game designs, 3D models, and data visualizations from different AI models. Participants can then compare results and provide valuable feedback, creating a rich dataset of user preferences and model performance.

Key Findings from Thousands of User Votes

With nearly 4,000 votes and around 5,000 users engaged, here are the standout observations:

Top Performing Models for Coding and UI/UX

Claude and DeepSeek Lead the Pack

Among the top contenders, models from the Claude series and DeepSeek have garnered the highest user preference ratings. Specifically, Claude Opus consistently ranked at the top, thanks to its strong performance in creating functional and aesthetically pleasing interfaces. DeepSeek models, especially version 0, excel in generating website content, making them valuable tools despite some limitations.

Speed Is a Consideration

While DeepSeek models are impressive, they tend to be slower in output, which might impact workflow speed. Therefore, if rapid interface development is a priority, Claude models could be the better choice.

Spotlight on an Underappreciated Model: Grok 3

Despite less online buzz—possibly due to its association with Elon Musk—Grok 3 surprises many by ranking within the top five. Notably, it delivers faster performance than many contemporaries, making it an underrated gem for developers seeking efficiency.

Mixed Results with Gemini 2.5-Pro

The Gemini 2.5-Pro model presents a mixed picture. Some users report excellent UI/UX generation capabilities, while others criticize its tendency to produce poorly designed applications. Interestingly, it performs reasonably well in coding business logic, indicating potential depending on the use case.

Relative Position of Leading Competitors

Open