Inefficient ventilatory response during cardiopulmonary exercise testing (CPET) has been suggested as a cause of post-COVID-19 dyspnea. It has been described in hospitalized patients (HOSP) with lung parenchymal sequelae but also after mild infection in ambulatory patients (AMBU). We hypothesize that AMBU and HOSP have different ventilatory responses to exercise, due to different etiologies. We analyzed CPET realized between July 2020 and May 2022 of patients with persisting respiratory symptoms 3 mo after COVID-19. Chest computed tomography (CT) scan, pulmonary function tests, quality of life, and respiratory questionnaires were collected. CPET data were specifically explored as a function of ventilation (V̇e) and time. Seventy-nine consecutive patients were included (42 AMBU and 37 HOSP, median: 54 [44-60] yr old, 57% female). Patients were hospitalized for a median of 20 [8-34] days, with pneumonia (41%) or acute respiratory distress syndrome (ARDS; 30%). Among HOSP, 12(32%) patients had abnormal values for spirometry and 18(51%) for carbon monoxide diffusing capacity (P< 0.001). CPET showed no differences between AMBU and HOSP in peak absolute O₂uptake (V̇o₂) (1.59 [1.22-2.11] mL·min^-1;P= 0.65). Tidal volume (VT) as a function of V̇e, was lower in AMBU than in HOSP (P< 0.01) toward the end of exercise. The slope of the V̇e-CO₂production was higher than normal in both groups (30.9 [26.1-34.3];P= 0.96). In conclusion, the severity of COVID-19 did not influence the exercise capacity, but AMBU demonstrated a less efficient ventilatory response to exercise as compared with HOSP. CPET with exploration of data as a function of V̇e and throughout the exercise better unveil ventilatory inefficiency. NEW & NOTEWORTHY We evaluated the exercise ventilatory response in patients with persisting dyspnea after severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2) infection. We found that despite similar peak power and peak absolute O₂uptake, tidal volume as a function of ventilation was lower in ambulatory than in hospitalized patients toward the end of exercise, reflecting ventilatory inefficiency. We call for evaluation of minute ventilation with the exploration of data throughout the exercise and not only peak data to better unveil ventilatory inefficiency.