Agent Startup Bench

Agent Startup Bench measures AI agents on high-economic-value, startup-style tasks that require autonomous planning and execution to deliver practical, verifiable results.

Released

Unknown

Readiness

Paper only

Primary field

General AI

Benchmark Radar records only publicly supported details and links back to primary sources for verification.