ABSTRACT Background: Sober living houses are designed for individuals in recovery to live with others in recovery, yet no guidelines exist for the time needed in a sober living house to significantly impact outcomes. Objectives: To examine how the length of stay in sober living houses is related to substance use and related outcomes, focusing on early discontinuation (length of stay less than six months) and stable residence (length of stay six months or longer). Methods: Baseline and 12-month data were collected from 455 sober living house residents (36% female). Longitudinal mixed models tested associations between early discontinuation vs. stable residence and abstinence, recovery capital, psychiatric, and legal outcomes. Final models were adjusted for resident demographics, treatment, 12-step attendance, use in social network, and psychiatric symptoms, with a random effect for house. Results: Both early discontinuers (n = 284) and stable residents (n = 171) improved significantly (Ps ≤ .05) between baseline and 12 months on all outcomes. Compared to early discontinuation, stable residence was related to 7.76% points more percent days abstinent (95% CI: 4.21, 11.31); 0.88 times fewer psychiatric symptoms (95% CI: 0.81, 0.94); 0.84 times fewer depression symptoms (95% CI: 0.76, 0.92); and lower odds of any DSM-SUD (OR = 0.65, 95% CI: 0.47, 0.89) and any legal problems (OR = 0.58, 95% CI: 0.40, 0.86). Conclusion: In this study of sober living houses in California, staying in a sober living house for at least six months was related to better outcomes than leaving before six months. Residents and providers should consider this in long-term recovery planning.