C 3 photosynthesis at high light is often modeled by assuming limitation by the maximum capacity of Rubisco carboxylation (V Cmax ) at low CO 2 concentrations, by electron transport capacity (J max ) at higher CO 2 concentrations, and sometimes by triose-phosphate utilization rate at the highest CO 2 concentrations. Net photosynthetic rate (P N ) at lower light is often modeled simply by assuming that it becomes limited by electron transport (J). However, it is known that Rubisco can become deactivated at less than saturating light, and it is possible that P N at low light could be limited by the rate of Rubisco carboxylation (V C ) rather than J. This could have important consequences for responses of P N to CO 2 and temperature at low light. In this work, P N responses to CO 2 concentration of common bean, quinoa, and soybean leaves measured over a wide range of temperatures and PPFDs were compared with rates modeled assuming either V C or J limitation at limiting light. In all cases, observed rates of P N were better predicted by assuming limitation by V C rather than J at limiting light both below and above the current ambient CO 2 . One manifestation of this plant response was that the relative stimulation of P N with increasing the ambient CO 2 concentration from 380 to 570 mol mol -1 did not decrease at less than saturating PPFDs. The ratio of V C to V Cmax at each lower PPFD varied linearly with the ratio of P N at low PPFD to P N at high PPFD measured at 380 mol mol -1 CO 2 in all cases. This modification of the standard C 3 biochemical model was much better at reproducing observed responses of light-limited P N to CO 2 concentrations from pre-industrial to projected future atmospheric concentrations.