Sensitivity of a Weibull distribution based model of complex-system reliability is investigated analytically and numerically. The complex system modeled is a central processing unit product. The types of sensitivity examined are that of the reliability model to erroneous parameter estimates, to improvements or degradations in component performance, and to variations in equipment utilization. In each case, the focus is upon the early life of the equipment. It is found that only the equipment utilization variations produce substantial model sensitivity.