The fate of plastics that enter the ocean is a longstanding puzzle. Recent estimates of the oceanic input of plastic are one to two orders of magnitude larger than the amount measured floating at the surface. This discrepancy could be due to overestimation of input estimates, processes removing plastic from the surface ocean or fragmentation and degradation. Here we present a 3D global marine mass budget of buoyant plastics that resolves this discrepancy. We assimilate observational data from different marine reservoirs, including coastlines, the ocean surface, and the deep ocean, into a numerical model, considering particle sizes of 0.1–1,600.0 mm. We find that larger plastics (>25 mm) contribute to more than 95% of the initially buoyant marine plastic mass: 3,100 out of 3,200 kilotonnes for the year 2020. Our model estimates an ocean plastic input of about 500 kilotonnes per year, less than previous estimates. Together, our estimated total amount and annual input of buoyant marine plastic litter suggest there is no missing sink of marine plastic pollution. The results support higher residence times of plastics in the marine environment compared with previous model studies, in line with observational evidence. Long-lived plastic pollution in the world’s oceans, which our model suggests is continuing to increase, could negatively impact ecosystems without countermeasures and prevention strategies.