Engineering track · Leaderboard
UI Design
Build complete, polished interfaces from a written brief. Scored on rendered output.
suite v1.0Run gemini-3.8-flash · published Sep 21, 2026
Ranking
Score is the mean of task medians (0-100). Consistency is the mean spread across repeats, lower is better. Cost per point is cost per task divided by score.
| 1 | Gemini 3.8 Flash google/gemini-3.8-flashGoogle · effort high | 87.1 | 1.3 | $0.085 | $0.00097 | 2m 10s | 0 |
Cost versus score
Each point is a model. Right is more expensive per task (log scale); up is better. The leader is highlighted.
Tasks in UI Design
Briefs and rubrics are public. Hidden tests, reference answers and private fixtures are not.