
Personal Agent Bench
Independent tests and teardowns of personal AI agents
Personal Agent Bench provides an independent benchmark for personal AI assistants that handle tasks such as booking, cancelling, and scheduling. It disassembles each agent at runtime, examining model execution, tool integration, memory storage, and user approvals, then subjects it to a repeatable task exam. Completion rates include a 95% confidence interval and sample size; the benchmark offers no aggregate score or ranking. The Zentor team, which develops agent software, excludes its own products from evaluation.