In high-risk for psychosis (CHR-P) much focus has been on the transition to psychosis. However, the desired outcome remains symptomatic remission and improved functioning as early as possible. This study aims to investigate predictors of sub-groups of remission and functional outcomes.A two-year prospective follow-up study of 104 CHR-P participants recruited in Norway using systematic early detection strategies. The Structural Interview for Prodromal Syndromes (SIPS) was used to assess CHR-P. Participants were classified as remitted or non-remitted based on their symptom scores. A latent class analysis was performed on the dichotomous data to identify latent classes regarding the timing of remission. t-tests and chi-squared tests were used to assess the association between class affiliations, predictors and outcomes.The latent class analysis showed moderate fit and divided the participants into three remission classes: "poor chance of remission" (16.7 %), "later remission" (34.3 %), and "early remission" (49.0 %). The "early remission" class had the highest probability of fast and stable remission, and had better premorbid academic, and baseline and 2-year global functioning than the "later remission" class. Baseline predictors such as age, SIPS symptoms, drug use, years in school and gender were not significantly associated with remission class.The study's main finding is the division of CHR-P remission into "early remission" and "later remission" and predictors of class affiliation. The monthly follow-up during the first six months allowed for the detection of this division. The findings suggest the importance of considering functioning in models of remission from CHR-P.