Cognitive behavioral therapy for insomnia (CBT-I) is a treatment with moderate to large effects. These effects are believed to be sustained long-term, but no systematic meta-analyses of recent evidence exist. In this present meta-analysis, we investigate long-term effects in 30 randomized controlled trials (RCTs) comparing CBT-I to non-active control groups. The primary analyses (n = 29 after excluding one study which was an outlier) showed that CBT-I is effective at 3-, 6- and 12-mo compared to non-active controls: Hedges g for Insomnia severity index: 0.64 (3 m), 0.40 (6 m) and 0.25 (12 m); sleep onset latency: 0.38 (3 m), 0.29 (6 m) and 0.40 (12 m); sleep efficiency: 0.51 (3 m), 0.32 (6 m) and 0.35 (12 m). We demonstrate that although effects decline over time, CBT-I produces clinically significant effects that last up to a year after therapy.