The impact of COVID-19 severity on development of long-term sequelae remains unclear, and symptom courses are not well defined. This ambidirectional cohort study recruited adults with new or worsening symptoms lasting ≥3 weeks from confirmed SARS-CoV-2 infection between August 2020-December 2021. COVID-19 severity was defined as severe for those requiring hospitalization and mild for those not. Symptoms were collected using standardized questionnaires. Multivariable logistical regression estimated odds ratios (OR) and 95% confidence intervals (CI) for associations between clinical variables and symptoms. Of 332 participants enrolled, median age was 52 years (IQR 42-62), 233 (70%) were female, and 172 (52%) were African American. Antecedent COVID-19 was mild in 171 (52%) and severe in 161 (48%). In adjusted models relative to severe cases, mild COVID-19 was associated with greater odds of fatigue (OR:1.83, CI:1.01-3.31), subjective cognitive impairment (OR:2.76, CI:1.53-5.00), headaches (OR:2.15, CI:1.05-4.44), and dizziness (OR:2.41, CI:1.18-4.92). Remdesivir treatment was associated with less fatigue (OR:0.47, CI:0.26-0.86) and fewer participants scoring >1.5 SD on PROMIS Cognitive scales (OR:0.43, CI:0.20-0.92). Fatigue and subjective cognitive impairment prevalence was higher 3-6 months after COVID-19 and persisted (fatigue OR:3.29, CI:2.08-5.20; cognitive OR:2.62, CI:1.67-4.11). Headache was highest at 9-12 months (OR:5.80, CI:1.94-17.3). Mild antecedent COVID-19 was associated with highly prevalent symptoms, and those treated with remdesivir developed less fatigue and cognitive impairment. Sequelae had a delayed peak, ranging 3-12 months post infection, and many did not improve over time, underscoring the importance of targeted preventative measures.