Human Preference Rankings for 60 Vibe-Coded Web Applications

60 web applications from the Vibe Coding Showcase were evaluated in a pairwise human preference study. The dataset contains 1,770 pairwise comparisons, with 30 human votes collected for each pair to judge visual design based on screenshots. It was created by datapointai and last updated in March 2026.

Use Cases

• Train a reward model to predict human design preference scores from application screenshots.
• Benchmark the correlation between automated UI metrics and human judgments from the 30-vote comparisons.
• Analyze consensus and disagreement patterns within the 1,770 pairwise evaluations.
• Fine-tune a vision-language model to generate design critiques using the pairwise preference labels as a signal.

Strengths

• 1,770 exhaustive pairwise comparisons covering all possible pairs of 60 apps.
• 30 human votes per comparison, providing a measure of consensus for each pair.

Limitations

• Limited to 60 applications from a single showcase, which may not represent the full diversity of web design.
• The visual design judgments are subjective and may reflect the specific preferences of the annotator pool.
• No raw application code or interactive elements are included, only screenshots and comparison results.

Provenance

Source

Vibe Coding Showcase, curated by datapointai.

Collection Method

Human annotators performed pairwise comparisons of application screenshots.

Freshness

Last updated March 2026.