In dictating the relative distances between the elbow, shoulder and wrist, avian brachial index (BI = humerus/radius‐ulna length) likely influences wing kinematics and, therefore, might predict extinct avian flight capability. This underpins the hypothesis that non‐neornithine Mesozoic avialans with relatively low BIs (associated with improved flight capabilities) restricted neornithine diversification until after the Cretaceous–Paleogene boundary. Here, correlations between flight metrics (wingbeat frequency (f), stroke angle (θ), wing loading (Q) and aspect ratio) and BI were investigated and vice versa. Additionally, the evolutionary model best describing the phylogenetic distribution of BI, and the temporal patterns in BI, flight metrics, body mass (Mb), and size‐corrected humerus (Lh) and radius‐ulna (Lru) length were determined. BI was best described by Ornstein–Uhlenbeck processes, with low α values indicating a gradual shift towards a future theoretical optimum. BI also decreased overall through evolutionary time with the flight metrics mirroring temporal patterns of change in BI. Mb, Lh and Lru overall decreased apart from increases in Lh and Lru following the middle‐late Miocene (also leading to BI increasing) due to diversifications of the Anatinae and Sphenisciformes. Lh overall decreased further than Lru. Consequently, decreasing Lh mainly contributed to decreasing BI through evolutionary time, implying flight performance increased through neornithine evolution. However, the timings of radiations in these variables implies an Eocene radiation of neornthine flight ecology rather than a rapid expansion into niches vacated by non‐neornithine Mesozoic avialians following the Cretaceous–Paleogene boundary. Multiple regressions showed f, θ and Q explained 60% of variation in BI. However, unequivocally evaluating whether BI is related to wing movement (and flight capability) requires direct measures of wing movement for many species, which are currently unavailable. Finally, the findings here and previously observed clade‐specificity in BI, suggest flight ecology may also be clade‐specific. Hence, the utility of phylogeny in predicting flight ecology requires exploration.