In model-based testing, software test-cases are derived from a formal specification. A popular technique is to use traces created by a model-checker as test-cases. This ap- proach is fully automated and flexible with regard to the structure and type of test-cases. Nondeterministic models, however, pose a problem to testing with model-checkers. Even though a model-checker is able to cope with nonde- terminism, the traces it returns make commitments at non- deterministic transitions. If a resulting test-case is executed on an implementation that takes a different, valid transition at such a nondeterministic choice, then the test-case would erroneously detect a fault. This paper discusses the exten- sion of available model-checker based test-case generation methods so that the problem of nondeterminism can be over- come.