Size is an important aspect of dinosaur paleontology, of interest to both the general public and professional scientists. Dinosaurs show some of the most extreme variations in size of any land animal group, ranging from the bee hummingbird, weighing just three grams, to the extinct titanosaurs, such as Argentinosaurus and Bruhathkayosaurus, which could weigh as much as 50–130 t (55–143 short tons).

The latest evidence suggests that dinosaurs' average size varied through the Triassic, early Jurassic, late Jurassic and Cretaceous periods, and dinosaurs probably only became widespread during the early or mid Jurassic. Predatory theropod dinosaurs, which occupied most terrestrial carnivore niches during the Mesozoic, most often fall into the 100–1,000 kg (220–2,200 lb) category when sorted by estimated weight into categories based on order of magnitude, whereas recent predatory carnivoran mammals peak in the range of 10–100 kg (22–220 lb). The mode of Mesozoic dinosaur body masses is between one and ten metric tonnes. This contrasts sharply with the size of Cenozoic mammals, estimated by the National Museum of Natural History as about 2 to 5 kg (4.4 to 11.0 lb).

When British paleontologist Richard Owen first coined the name Dinosauria in 1842, he made note of the size of the "gigantic Crocodile-lizards of dry land". He believed that Megalosaurus was 9.1 m (30 ft) long and built like large modern mammals, larger than any other known saurian.

Scientists will probably never be certain of the largest and smallest dinosaurs. This is because only a small fraction of animals ever fossilize, and most of these remains will likely never be uncovered. Of the specimens that are recovered, few are even relatively complete skeletons, and impressions of skin and other soft tissues are rarely discovered. Rebuilding a complete skeleton by comparing the size and morphology of bones to those of similar, better-known species is an inexact art (though governed by some established allometric trends), and reconstructing the muscles and other organs of the living animal is, at best, a process of educated guesswork, and never perfect. Mass estimates for dinosaurs are much more variable than length estimates given the lack of soft tissue preservation in the fossilization process. Modern mass estimation is often done with the laser scan skeleton technique that puts a "virtual" skin over the known or implied skeleton, but the limitations inherent in previous mass estimation techniques remain.

Sauropodomorph size is difficult to estimate given their usually fragmentary state of preservation. Sauropods are often preserved without their tails, so the margin of error in overall length estimates is high. Mass is calculated using the cube of the length, so for species in which the length is particularly uncertain, the weight is even more so. Estimates that are particularly uncertain (due to very fragmentary or lost material) are preceded by a question mark. Each number represents the highest estimate of a given research paper. One large sauropod, Maraapunisaurus fragillimus, was based on particularly scant remains that have been lost since their description by paleontologists in 1878. Analysis of the illustrations included in the original report suggested that M. fragillimus may have been the largest land animal of all time, possibly weighing 100–150 t (110–170 short tons) and measuring between 40–60 m (130–200 ft) long. One later analysis of the surviving evidence, and the biological plausibility of such a large land animal, suggested that the enormous size of this animal was an over-estimate due partly to typographical errors in the original report. This would later be challenged by a different study, which argued Cope's measurements were genuine and that there was no basis for assuming typographical errors. The study, however, also reclassified the species and correspondingly gave a much lower length estimate of 30.3 metres (99 ft) and a mass of 78.5 t (86.5 short tons). This in itself would later be disputed as being too small for an animal of such size, with some believing it to be even larger at around 35–40 metres (115–131 ft) and weighing around 80–120 t (88–132 short tons).

Another large but even more controversial sauropod is Bruhathkayosaurus, which had a calculated weight ranging between 126–220 t (139–243 short tons) and a length of 44.1 m (145 ft) Although the existence of this sauropod had long been dismissed as a potential fake or a misidentification of a petrified tree trunk, recent photographic evidence emerged, confirming its existence. More recent and reliable estimates in 2023 have rescaled Bruhathkayosaurus to weigh around 110–130 t (120–140 short tons) with its most liberal estimate being 240 t (260 short tons), making it incredibly massive for such an animal. If the upper unlikely size estimates were to be taken at face value, Bruhathkayosaurus would not only be the largest dinosaur to have ever lived, but also the largest animal to have lived, exceeding even the largest blue whale recorded. According to Gregory S. Paul, 'super-sauropods' or 'land-whales' such as Maraapunisaurus, Bruhathkayosaurus and the "Broome Titanosaur footprints," as he calls them, should not be surprising as sauropods were more heat tolerant and grew rapidly, which allowed them to reach truly titanic sizes that rivalled the largest whales in mass despite the prevalence of air sacs. Moreover, in a 2024 publication, Gregory S. Paul further argued that a sauropod's internal air sacs were not as dramatically density reducing as has been widely thought, suggesting that masses reaching close to 200 tons was within the realm of possibility. In fact, a sauropod's r-selection reproduction combined with the positive feedback loop of needing longer necks to feed a growing fermentation factory of its digestive organs, allowed adult sauropods to balloon up in size when compared to their mammalian contemporaries, as a mammal's K-selected reproduction would have limited adult sizes to avoid the overexploitation of their habitat's food resources, something sauropods lacked entirely. Other potential factors for such extreme sauropod sizes include increasing bone robustness and load-distributing cartilaginous features to better redistribute and support such massive weights.

Generally, the giant sauropods can be divided into two categories: the shorter but stockier and more massive forms (mainly titanosaurs and some brachiosaurids), and the longer but slenderer and more light-weight forms (mainly diplodocids).

Because different methods of estimation sometimes give conflicting results, mass estimates for sauropods can vary widely causing disagreement among scientists over the accurate number. For example, the titanosaur Dreadnoughtus was originally estimated to weigh 59.3 tonnes by the allometric scaling of limb-bone proportions, whereas more recent estimates, based on three-dimensional reconstructions, yield a much smaller figure of 22.1–38.2 tonnes.