Benchmark
App-Bench
A benchmark for evaluating AI systems that generate modern web applications.
AfterQuery · 100% completeCompany-reported
Environments where agents produce and refine visual work (interfaces, layouts, graphics) with reward signals from human taste panels and programmatic design checks. Design is hard to verify automatically, making it a frontier for rubric-based and preference-based rewards.
1 company · 1 cataloged artifact · as of 2026-08-16
| Name | Domains | Type / catalog | Location | Team | Funding | Raising |
|---|---|---|---|---|---|---|
| Taste Labs Design-domain environments and evals for AI models | Design | Pure-play commercial Catalog unknown | New York | 26-50 | - | Unknown |
Source-backed catalog records connected to design.