Present paper inspects the prediction capability of the latest version of the International Reference Ionosphere (IRI‐2012) model in predicting the total electron content (TEC) over seven different equatorial regions across the globe during a very low solar activity phase 2009 and a high solar activity phase 2012. This has been carried out by comparing the ground‐based Global Positioning System (GPS)‐derived VTEC with those from the IRI‐2012 model. The observed GPS‐TEC shows the presence of winter anomaly which is prominent during the solar maximum year 2012 and disappeared during solar minimum year 2009. The monthly and seasonal mean of the IRI‐2012 model TEC with IRI‐NeQ topside has been compared with the GPS‐TEC, and our results showed that the monthly and seasonal mean value of the IRI‐2012 model overestimates the observed GPS‐TEC at all the equatorial stations. The discrepancy (or over estimation) in the IRI‐2012 model is found larger during solar maximum year 2012 than that during solar minimum year 2009. This is a contradiction to the results recently presented by Tariku (2015) over equatorial regions of Uganda. The discrepancy is found maximum during the December solstice and a minimum during the March equinox. The magnitude of discrepancy in the IRI‐2012 model showed longitudinal dependent which maximized in western longitude sector during both the years 2009 and 2012. The significant discrepancy in the IRI‐2012 model observed during the solar minimum year 2009 could be attributed to larger difference between F10.7 flux and EUV flux (26–34 nm) during low solar activity period 2007–2009 than that during high solar activity period 2010–2012. This suggests that to represent the solar activity impact in the IRI model, implementation of new solar activity indices is further required for its better performance.