Malaria continues to kill approximately 650 000 people each year. There is evidence that some second-generation insecticide-treated nets, which combine insecticide formulations with different modes of action, are protective against malaria while the nets are new; however, evidence for their impact over 3 years is scarce. In this study, we report the third-year results of a cluster-randomised controlled trial assessing the long-term effectiveness of dual-active ingredient long-lasting insecticidal nets (LLINs). This is a secondary analysis of a cluster-randomised controlled trial, carried out between May 23, 2019, and April 30, 2023, in southern Benin. Restricted randomisation was used to assign 60 clusters (villages or groups of villages with a minimum of 100 households) to the three study groups (1:1:1) to evaluate the efficacy of pyriproxyfen-pyrethroid LLINs and chlorfenapyr-pyrethroid LLINs compared with pyrethroid-only LLINs (reference) against malaria transmission. The study staff and communities were masked to the group allocation. The primary outcome was malaria incidence measured over the third year after LLIN distribution, in a cohort of children aged 6 months to 9 years at the time of enrolment, in the intention-to-treat population. Here, we present the data of the third year post-LLIN distribution. The trial was registered with ClinicalTrials.gov, NCT03931473. Study net use declined over the 3 years and was consistently lowest in the pyriproxyfen-pyrethroid LLIN group (at 36 months: 889 [39·4%] of 2257 participants vs 1278 [52·2%] of 2450 participants for the chlorfenapyr-pyrethroid LLIN group and 1400 [57·6%] of 2430 participants for the pyrethroid-only LLIN group). The cohort of children for the third year of follow-up (600 per group) were enrolled between April 9 and 30, 2022. Mean malaria incidence during the third year after distribution was 1·19 cases per child-year (95% CI 1·09-1·29) in the pyrethroid-only LLIN reference group, 1·21 cases per child-year (1·12-1·31) in the pyriproxyfen-pyrethroid LLIN group (hazard ratio [HR] 1·02, 95% CI 0·71-1·44; p=0·92), and 0·96 cases per child-year (0·88-1·05) in the chlorfenapyr-pyrethroid LLIN group (HR 0·80, 0·56-1·17; p=0·25). No adverse events related to study nets were reported by participants. During the third year, as was also observed during the first 2 years, the pyriproxyfen-pyrethroid LLIN group did not have superior protection against malaria cases compared with the standard LLIN group. In the third year, people living in the chlorfenapyr-pyrethroid LLIN group no longer benefited from greater protection against malaria cases and infections than those living in the pyrethroid-only LLIN group. This was probably influenced by lower study net use than previous years and the declining concentration of partner insecticides in the nets. UNITAID, The Global Fund. For the French translation of the abstract see Supplementary Materials section.