Twelve existing lifetime models for optical fibers, based on a power law to describe the stress- induced crack growth, are studied and compared in COST-218, WG1. The number of models is reduced to only 1 basic model, taking into account the effects of proof test. An alternative model, to be used for proof testing on-line, is given as a worst-case limit. A choice of 3 testing methods to obtain information about the weak-flaw distribution is given: dynamic- fatigue and variable screen-testing of long lengths or using the failure number during proof test. The presented model can also be worked up for different (more harsh) environments during service, without the need to know the corrosion parameter B, provided that stress free aging does not occur. It is explained which additional measurements in environments as during service and (proof) tests have to be performed. The presented model, based upon weak flaw statistics, is also shown to be useful in many short length applications, e.g., splice enclosures and optical amplifiers.