The boron isotope (δ¹¹B) proxy for seawater pH is a tried and tested means to reconstruct atmospheric CO₂ in the geologic past, but uncertainty remains over how to treat species-specific calibrations that link foraminiferal δ¹¹B to pH estimates prior to 22 My. In addition, no δ¹¹B-based reconstructions of atmospheric CO₂ exist for wide swaths of the Oligocene (33–23 Ma), and large variability in CO₂ reconstructions during this epoch based on other proxy evidence leaves climate evolution during this period relatively unconstrained. To add to our understanding of Oligocene and early Miocene climate, we generated new atmospheric CO₂ estimates from new δ¹¹B data from fossil shells of surface-dwelling planktic foraminifera from the mid-Oligocene to early Miocene (∼28–18 Ma). We estimate atmospheric CO₂ of ∼680 ppm for the mid-Oligocene, which then evolves to fluctuate between ∼500–570 ppm during the late Oligocene and between ∼420–700 ppm in the early Miocene. These estimates tend to trend higher than Oligo-Miocene CO₂ estimates from other proxies, although we observe good proxy agreement in the late Oligocene. Reconstructions of CO₂ fall lower than estimates from paleoclimate model simulations in the early Miocene and mid Oligocene, which indicates that more proxy and/or model refinement is needed for these periods. Our species cross-calibrations, assessing δ¹¹B, Mg/Ca, δ¹⁸O, and δ¹³C, are able to pinpoint and evaluate small differences in the geochemistry of surface-dwelling planktic foraminifera, lending confidence to paleoceanographers applying this approach even further back in time.