Background and aims: Insulin resistance (IR) is a risk factor for several cardiometabolic disorders; however, there is conflicting evidence about the reliability of certain IR markers. In this context, the triglyceride-glucose index (TyG) has been proposed as a surrogate marker for IR. This study aimed to compare the TyG index and homeostasis model assessment of insulin resistance (HOMA-IR). Methods and results: A cross-sectional analysis was conducted using baseline data from 11,314 adults (aged 35–74 years) from the ELSA-Brasil study. The correlation between TyG and HOMA-IR, their interrater reliability, and their predictive value in identifying metabolic syndrome (MetS) were assessed. The mean TyG and HOMA-IR in our sample were 8.81 ± 0.52 and 2.78 ± 1.58 for men, and 8.53 ± 0.48 and 2.49 ± 1.38 for women, respectively. TyG and HOMA-IR showed a weak to moderate correlation with each other (Pearson’s r for men: 0.395 and 0.409 for women, p-value <0.05) and other markers of glycemic metabolism. Additionally, the area under the curve for the prediction of MetS was greater for TyG than HOMA-IR, regardless of sex (TyG: 0.836 for men and 0.826 for women; HOMA-IR: 0.775 for men and 0.787 for women). The concordance between these markers was low (Coheńs kappa coefficient: 0.307 for men and 0.306 for women). Individuals with increased TyG exhibited mainly anthropometrical and glycemic metabolic alterations, whereas those with elevated HOMA-IR displayed mostly lipid-associated metabolic alterations. Conclusion: TyG and HOMA-IR might indicate different profiles of cardiometabolic disorders, showing poor agreement in classifying individuals (normal vs. altered) and a weak correlation. Therefore, further studies are needed to investigate the role of TyG as a surrogate marker of IR.