Treatment-resistant schizophrenia (TRS) is an important and unresolved problem in biological and clinical psychiatry. Approximately 30% of cases of schizophrenia (Sch) are TRS, which may be due to the fact that some patients with TRS may suffer from pathogenetically “non-dopamine” Sch, in the development of which neuroinflammation is supposed to play an important role. The purpose of this narrative review is an attempt to summarize the data characterizing the patterns of production of pro-inflammatory and anti-inflammatory cytokines during the development of therapeutic resistance to APs and their pathogenetic and prognostic significance of cytokine imbalance as TRS biomarkers. This narrative review demonstrates that the problem of evaluating the contribution of pro-inflammatory and anti-inflammatory cytokines to maintaining or changing the cytokine balance can become a new key in unlocking the mystery of “non-dopamine” Sch and developing new therapeutic strategies for the treatment of TRS and psychosis in the setting of acute and chronic neuroinflammation. In addition, the inconsistency of the results of previous studies on the role of pro-inflammatory and anti-inflammatory cytokines indicates that the TRS biomarker, most likely, is not the serum level of one or more cytokines, but the cytokine balance. We have confirmed the hypothesis that cytokine imbalance is one of the most important TRS biomarkers. This hypothesis is partially supported by the variable response to immunomodulators in patients with TRS, which were prescribed without taking into account the cytokine balance of the relation between serum levels of the most important pro-inflammatory and anti-inflammatory cytokines for TRS.