Background Adolescent physical activity levels are low and are associated with rising disease risk and social disadvantage. The Get Others Active (GoActive) intervention was co-designed with adolescents and teachers to increase physical activity in adolescents. Objective To assess the effectiveness and cost-effectiveness of the school-based GoActive programme in increasing adolescents’ moderate-to-vigorous physical activity. Design A cluster randomised controlled trial with an embedded mixed-methods process evaluation. Setting Non-fee-paying schools in Cambridgeshire and Essex, UK (n = 16). Schools were computer randomised and stratified by socioeconomic position and county. Participants A total of 2862 Year 9 students (aged 13–14 years; 84% of eligible students). Intervention The iteratively developed feasibility-tested refined 12-week intervention trained older adolescents (mentors) and in-class peer leaders to encourage classes to undertake two new weekly activities. Mentors met with classes weekly. Students and classes gained points and rewards for activity in and out of school. Main outcome measures The primary outcome was average daily minutes of accelerometer-assessed moderate-to-vigorous physical activity at 10 months post intervention. Secondary outcomes included accelerometer-assessed activity during school, after school and at weekends; self-reported physical activity and psychosocial outcomes; cost-effectiveness; well-being and a mixed-methods process evaluation. Measurement staff were blinded to allocation. Results Of 2862 recruited participants, 2167 (76%) attended 10-month follow-up measurements and we analysed the primary outcome for 1874 (65.5%) participants. At 10 months, there was a mean decrease in moderate-to-vigorous physical activity of 8.3 (standard deviation 19.3) minutes in control participants and 10.4 (standard deviation 22.7) minutes in intervention participants (baseline-adjusted difference –1.91 minutes, 95% confidence interval –5.53 to 1.70 minutes; p = 0.316). The programme cost £13 per student compared with control. Therefore, it was not cost-effective. Non-significant indications of differential impacts suggested detrimental effects among boys (boys –3.44, 95% confidence interval –7.42 to 0.54; girls –0.20, 95% confidence interval –3.56 to 3.16), but favoured adolescents from lower socioeconomic backgrounds (medium/low 4.25, 95% confidence interval –0.66 to 9.16; high –2.72, 95% confidence interval –6.33 to 0.89). Mediation analysis did not support the use of any included intervention components to increase physical activity. Some may have potential for improving well-being. Students, teachers and mentors mostly reported enjoying the GoActive intervention (56%, 87% and 50%, respectively), but struggled to conceptualise their roles. Facilitators of implementation included school support, embedding a routine, and mentor and tutor support. Challenges to implementation included having limited school space for activities, time, and uncertainty of teacher and mentor roles. Limitations Retention on the primary outcome at 10-month follow-up was low (65.5%), but we achieved our intended sample size, with retention comparable to similar trials. Conclusions A rigorously developed school-based intervention (i.e. GoActive) was not effective in countering the age-related decline in adolescent physical activity. Overall, this mixed-methods evaluation provides transferable insights for future intervention development, implementation and evaluation. Future work Interdisciplinary research is required to understand educational setting-specific implementation challenges. School leaders and authorities should be realistic about expectations of the effect of school-based physical activity promotion strategies implemented at scale. Trial registration Current Controlled Trials ISRCTN31583496. Funding This project was funded by the National Institute for Health Research (NIHR) Public Health Research programme and will be published in full in Public Health Research; Vol. 9, No. 6. See the NIHR Journals Library website for further project information. This work was additionally supported by the Medical Research Council (London, UK) (Unit Programme number MC_UU_12015/7) and undertaken under the auspices of the Centre for Diet and Activity Research (Cambridge, UK), a UK Clinical Research Collaboration Public Health Research Centre of Excellence. Funding from the British Heart Foundation (London, UK), Cancer Research UK (London, UK), Economic and Social Research Council (Swindon, UK), Medical Research Council, the National Institute for Health Research (Southampton, UK) and the Wellcome Trust (London, UK), under the auspices of the UK Clinical Research Collaboration, is gratefully acknowledged (087636/Z/08/Z; ES/G007462/1; MR/K023187/1). GoActive facilitator costs were borne by Essex and Cambridgeshire County Councils.