Odontoceti (toothed whales) and Mysticeti (baleen whales) diverged ~39 Ma. Odontocetes evolved high-frequency echolocation and shifted cranial bones posteriorly, while mysticetes evolved larger masses and filter-feeding. Despite an excellent fossil record, little quantitative study of shape evolution spanning cetacean diversity has been done. Here, we quantify morphological disparity and evolutionary rate in cranial shape to identify the major factors driving its evolution, including ecological traits and palaeoclimate. We scanned 88 living and 111 Eocene to Pliocene fossil cetaceans, representing the largest 3D dataset for cetaceans ever constructed. We used 123 3D coordinate-based landmarks and 1905 curve semilandmarks to capture cranial shape and analyzed data within a phylogenetic framework. The largest component of cranial variation (PC1 = 41.6%) reflects a posterior shift in the nares, followed by rostrum length (PC2 = 23.0%) with dolichocephalic (e.g., Pontoporia blainvillei) and brachycephalic (e.g., Kogia sima) crania representing the extremes. After accounting for phylogeny, habitat had no significant effect on skull shape, but diet, feeding, and tooth type did. The highest morphological disparity is in the premaxilla and nasal, with the highest evolutionary rate ( 2 mult) in the frontal. Disparity rises rapidly in the Middle-Late Oligocene and peaks in the Middle-Late Miocene. A model with rate of cranial evolution tracking paleotemperature outperformed Brownian motion and Early Burst models and suggests highest rates in the Latest Eocene and Early-Mid Oligocene (during periods of global cooling) with a further peak in evolutionary rates during the Mid-Miocene warming period.