Background Studying individuals with varying symptoms, from mild to severe, can provide valuable insights into the spectrum of cognitive outcomes after COVID-19. We investigated the cognitive performance of adults who recovered from the novel coronavirus disease (COVID-19) without prior cognitive complaints, considering mild (not hospitalized), moderate (ward), and severe (intensive care unit) symptoms. Methods This cross-sectional study included 302 patients who recovered from COVID-19 (mild, n = 102; moderate, n = 102; severe, n = 98). We assessed intellectual quotient (IQ), attention, memory, processing speed, visual-constructive ability, as well as symptoms of depression, anxiety, and stress, at least eighteen months after infection. The mean length of hospitalization was Mdays=8.2 (SD = 3.9) and Mdays=14.4 (SD = 8.2) in the moderate and severe groups, respectively. Results Cognitive difficulties were present in all three groups: mild (n = 12, 11.7%), moderate (n = 40, 39.2%), and severe (n = 48, 48.9%). Using Multinomial Logistic Regression and considering the odds ratio, our results indicated that a one-point increase in sustained attention, visual memory, and working memory might decrease the odds of being categorized in the severe group by 20%, 24%, and 77%, respectively, compared to the mild group. Conclusions Our findings provide empirical evidence regarding the long-term cognitive effects of COVID-19, particularly in individuals experiencing severe manifestations of the disease. We also highlighted the need for a comprehensive, multidimensional approach in rehabilitation programs to address the enduring cognitive impacts of COVID-19.