If long-term responses of photosynthesis and leaf diffusive conductance to rising atmospheric carbon dioxide (CO2) levels are similar or predictably different among species, functional types, and ecosystem types, general global models of elevated CO2 effects can effectively be developed. To address this issue we measured gas exchange rates of 13 perennial grassland species from four functional groups across 11 years of long-term free-air CO2 enrichment (eCO2, +180 ppm above ambient CO2) in the BioCON experiment in Minnesota, USA. Eleven years of eCO2 produced consistent but modest increases in leaf net photosynthetic rates of 10% on average compared with plants grown at ambient CO2 concentrations across the 13 species. This eCO2-induced enhancement did not depend on soil N treatment, is much less than the average across other longer-term studies, and represents strong acclimation (i.e. downregulation) as it is also much less than the instantaneous response to eCO2. The legume and C3 nonlegume forb species were the most responsive among the functional groups (+13% in each), the C4 grasses the least responsive (+4%), and C3 grasses intermediate in their photosynthetic response to eCO2 across years (+9%). Leaf stomatal conductance and nitrogen content declined comparably across species in eCO2 compared with ambient CO2 and to degrees corresponding to results from other studies. The significant acclimation of photosynthesis is explained in part by those eCO2-induced decreases in leaf N content and stomatal conductance that reduce leaf photosynthetic capacity in plants grown under elevated compared with ambient CO2 concentrations. Results of this study, probably the longest-term with the most species, suggest that carbon cycle models that assume and thereby simulate long-lived strong eCO2 stimulation of photosynthesis (e.g.> 25%) for all of Earth's terrestrial ecosystems should be viewed with a great deal of caution.