Benchmark
AI Deploy Bench
Evaluate the path from an application specification to a tested cloud deployment.
Explore the original evidence
Explore the original evaluation, its methods, and reported findings.
Open evalCoverage
Five app types across several cloud platforms, including sites, APIs, dashboards, and chat.
Methodology
Have an agent build and deploy each app, then check the live deployment independently. Record completion, human setup, time, tokens, cost, and security-review findings.