BACKGROUND: According to current consensus guidelines for type 2 diabetes management, bodyweight management is as important as attaining glycaemic targets. Retatrutide, a single peptide with agonist activity at the glucose-dependent insulinotropic polypeptide (GIP), GLP-1, and glucagon receptors, showed clinically meaningful glucose-lowering and bodyweight-lowering efficacy in a phase 1 study. We aimed to examine the efficacy and safety of retatrutide in people with type 2 diabetes across a range of doses. METHODS: In this randomised, double-blind, double-dummy, placebo-controlled and active comparator-controlled, parallel-group, phase 2 trial, participants were recruited from 42 research and health-care centres in the USA. Adults aged 18-75 years with type 2 diabetes, glycated haemoglobin (HbA) of 7·0-10·5% (53·0-91·3 mmol/mol), and BMI of 25-50 kg/m were eligible for enrolment. Eligible participants were treated with diet and exercise alone or with a stable dose of metformin (≥1000 mg once daily) for at least 3 months before the screening visit. Participants were randomly assigned (2:2:2:1:1:1:1:2) using an interactive web-response system, with stratification for baseline HbA and BMI, to receive once-weekly injections of placebo, 1·5 mg dulaglutide, or retatrutide maintenance doses of 0·5 mg, 4 mg (starting dose 2 mg), 4 mg (no escalation), 8 mg (starting dose 2 mg), 8 mg (starting dose 4 mg), or 12 mg (starting dose 2 mg). Participants, study site personnel, and investigators were masked to treatment allocation until after study end. The primary endpoint was change in HbA from baseline to 24 weeks, and secondary endpoints included change in HbA and bodyweight at 36 weeks. Efficacy was analysed in all randomly assigned, except inadvertently enrolled, participants, and safety was assessed in all participants who received at least one dose of study treatment. The study is registered at ClinicalTrials.gov, NCT04867785. FINDINGS: Between May 13, 2021, and June 13, 2022, 281 participants (mean age 56·2 years [SD 9·7], mean duration of diabetes 8·1 years [7·0], 156 [56%] female, and 235 [84%] White) were randomly assigned and included in the safety analysis (45 in the placebo group, 46 in the 1·5 mg dulaglutide group, and 47 in the retatrutide 0·5 mg group, 23 in the 4 mg escalation group, 24 in the 4 mg group, 26 in the 8 mg slow escalation group, 24 in the 8 mg fast escalation group, and 46 in the 12 mg escalation group). 275 participants were included in the efficacy analyses (one each in the retatrutide 0·5 mg group, 4 mg escalation group, and 8 mg slow escalation group, and three in the 12 mg escalation group were inadvertently enrolled). 237 (84%) participants completed the study and 222 (79%) completed study treatment. At 24 weeks, least-squares mean changes from baseline in HbA with retatrutide were -0·43% (SE 0·20; -4·68 mmol/mol [2·15]) for the 0·5 mg group, -1·39% (0·14; -15·24 mmol/mol [1·56]) for the 4 mg escalation group, -1·30% (0·22; -14·20 mmol/mol [2·44]) for the 4 mg group, -1·99% (0·15; -21·78 mmol/mol [1·60]) for the 8 mg slow escalation group, -1·88% (0·21; -20·52 mmol/mol [2·34]) for the 8 mg fast escalation group, and -2·02% (0·11; -22·07 mmol/mol [1·21]) for the 12 mg escalation group, versus -0·01% (0·21; -0·12 mmol/mol [2·27]) for the placebo group and -1·41% (0·12; -15·40 mmol/mol [1·29]) for the 1·5 mg dulaglutide group. HbA reductions with retatrutide were significantly greater (p<0·0001) than placebo in all but the 0·5 mg group and greater than 1·5 mg dulaglutide in the 8 mg slow escalation group (p=0·0019) and 12 mg escalation group (p=0·0002). Findings were consistent at 36 weeks. Bodyweight decreased dose dependently with retatrutide at 36 weeks by 3·19% (SE 0·61) for the 0·5 mg group, 7·92% (1·28) for the 4 mg escalation group, 10·37% (1·56) for the 4 mg group, 16·81% (1·59) for the 8 mg slow escalation group, 16·34% (1·65) for the 8 mg fast escalation group, and 16·94% (1·30) for the 12 mg escalation group, versus 3·00% (0·86) with placebo and 2·02% (0·72) with 1·5 mg dulaglutide. For retatrutide doses of 4 mg and greater, decreases in weight were significantly greater than with placebo (p=0·0017 for the 4 mg escalation group and p<0·0001 for others) and 1·5 mg dulaglutide (all p<0·0001). Mild-to-moderate gastrointestinal adverse events, including nausea, diarrhoea, vomiting, and constipation, were reported in 67 (35%) of 190 participants in the retatrutide groups (from six [13%] of 47 in the 0·5 mg group to 12 [50%] of 24 in the 8 mg fast escalation group), six (13%) of 45 participants in the placebo group, and 16 (35%) of 46 participants in the 1·5 mg dulaglutide group. There were no reports of severe hypoglycaemia and no deaths during the study. INTERPRETATION: In people with type 2 diabetes, retatrutide showed clinically meaningful improvements in glycaemic control and robust reductions in bodyweight, with a safety profile consistent with GLP-1 receptor agonists and GIP and GLP-1 receptor agonists. These phase 2 data also informed dose selection for the phase 3 programme. FUNDING: Eli Lilly and Company.