In this article we consider four versions of quantum mechanics (QM), which form four different theories and which have the same experimental consequences (one of them is the standard QM). We consider six fundamental problems as criteria for the evaluation of these theories. The most successful is the modified QM introduced in [1] and described in the axiomatic form in [2], which solves all six problems. The least successful is the standard QM, which does not solve any of considered problems. (It seems that the standard QM is unable to solve any of these problems.) Other two theories solve some of problems and did not solve some other problems. On the base of this result we recommend to abandon the standard QM and to accept the modified QM as the right QM, since both variants have the same experimental content and the differences rest only in theoretical considerations.
From the dichotomy nonlocality vs which is the consequence of Bell Inequalities (BI) we shall choose the non-realism. We shall present here the modified Quantum Mechanics (modQM) in the axiomatic form. ModQM was introduced in [5] and we shall show its non-realism in the description of an internal measurement process. ModQM allows the restoration of locality, since BI cannot be derived in it. In modQM it is possible to solve: the measurement problem, the collapse problem, the problem of a local model for EPR correlations (see[5]). ModQM is a unique explicit realization of non-realism in QM. ModQM should be preferred as an alternative to the standard QM mainly since it restores the locality.