Background In psychotic and mood disorders, immune alterations are hypothesized to underlie cognitive symptoms, as they have been associated with elevated blood levels of inflammatory cytokines, kynurenine metabolites, and markers of microglial activation. The current meta-analysis synthesizes all available clinical evidence on the associations between immunomarkers (IMs) and cognition in these psychiatric illnesses. Methods Pubmed, Web of Science, and Psycinfo were searched for peer-reviewed studies on schizophrenia spectrum disorder (SZ), bipolar disorder (BD), or major depressive disorder (MDD) including an association analysis between at least one baseline neuropsychological outcome measure (NP) and one IM (PROSPERO ID:CRD42021278371). Quality assessment was performed using BIOCROSS. Correlation meta-analyses, and random effect models, were conducted in Comprehensive Meta-Analysis version 3 investigating the association between eight cognitive domains and pro-inflammatory and anti-inflammatory indices (PII and AII) as well as individual IM. Results Seventy-five studies (n = 29,104) revealed global cognitive performance (GCP) to be very weakly associated to PII (r = −0.076; p = 0.003; I2 = 77.4) or AII (r = 0.067; p = 0.334; I2 = 38.0) in the combined patient sample. Very weak associations between blood–based immune markers and global or domain-specific GCP were found, either combined or stratified by diagnostic subgroup (GCP x PII: SZ: r = −0.036, p = 0.370, I2 = 70.4; BD: r = −0.095, p = 0.013, I2 = 44.0; MDD: r = −0.133, p = 0.040, I2 = 83.5). We found evidence of publication bias. Discussion There is evidence of only a weak association between blood-based immune markers and cognition in mood and psychotic disorders. Significant publication and reporting biases were observed and most likely underlie the inflation of such associations in individual studies.