Design
Sponsored byCheck whether an interface a model builds works for a real person: they find things, finish the task, and the layout holds up under real content. Or test whether a model's critique predicts what users actually do. Crowd votes reward how a page looks; the one benchmark built on real A/B results finds models weak.
Benchmarks to know
- Design Arena: Websites Design Arena
People vote blind between two sites built from the same prompt; the clearest public signal of web taste.
- Human Creativity Benchmark Contra Labs
Professional creatives judge landing pages, apps, ads and video in pairs; no model leads every phase in any domain.
- Design Arena: Slides Design Arena
Agents build a deck and people vote on its design; the gaps between models are wide and easy to see.
- Design Arena: SVG Design Arena
Blind votes on vector graphics from the same prompt: logos, icons and illustration.
- Vibe Code Bench Vals AI
Build a whole web app from a plain-English spec, tested by clicking through it; no model passes every test.
- UI/UX Image Arena Artificial Analysis
Blind votes on interface mockups from image models, where type and layout still give them away.
Bring an eval in this track, or in any other, on Saturday, December 5, 2026 at Cornell University.