Frontend Design Benchmark
Frontend design research
A redesigned gallery comparing AI-built websites across five shared briefs, with design skills on and off.
View benchmarkWhat we tested
The current gallery includes GPT 6 Sol, Luna, and Astra, Claude Opus 5, and Space Bunny Free across five product briefs, with and without design skills. Claude Opus 5.5 and Sonnet 5.5 are planned additions. Earlier GPT 5.6 and Claude Opus 4.8 results remain in the legacy collection.
What changed
The gallery received a full visual redesign, with a new model lineup, larger build previews, refreshed typography and color, and clearer controls for choosing models and design skills. The comparison viewer was also refined for mobile screens, full-size previews, and shareable selections.
How to explore it
Pick a brief to inspect its builds, switch the preview model and skill setting, or open the side-by-side comparison. Each build can also be opened on its own, and the earlier results are still accessible from the legacy section.