To determine the impact of seasonal steer grazing on annual CO₂ fluxes of annually burned native tallgrass prairie, we used relaxed eddy accumulation on adjacent pastures of grazed and ungrazed tallgrass prairie from 1998 to 2001. Fluxes of CO₂ were measured almost continuously from immediately following burning through the burn date the following year. Aboveground biomass and leaf area were determined by clipping biweekly during the growing season. Carbon lost because of burning was estimated by clipping immediately prior to burning. Soil CO₂ flux was measured biweekly each year using portable chambers. Steers were stocked at twice the normal season-long stocking rate (0.81 ha steer⁻¹) for the first half of the grazing season (∼ May 1 to July 15) and the area was left ungrazed the remainder of the year. That system of grazing is termed “intensive-early stocking.” During the early growing season, grazing reduced net carbon exchange relative to the reduction in green leaf area, but as the growing season progressed on the grazed area, regrowth produced younger leaves that had an apparent higher photosynthetic efficiency. Despite a substantially greater green leaf area on the ungrazed area, greater positive net carbon flux occurred on the grazed area during the late season. Net CO₂ exchange efficiency was greatest when grazing utilization was highest. We conclude that with grazing the reduced ecosystem respiration, the open canopy architecture, and the presence of young, highly photosynthetic leaves are responsible for the increased net carbon exchange efficiency. Both GR and UG tallgrass prairie appeared to be carbon-storage neutral for the 3 years of data collection (1998 ungrazed: −31 g C·m⁻², 1998 grazed: −5 g C·m⁻²; 1999 ungrazed: −40 g C·m⁻², 1999 grazed: −11 g C·m⁻²; 2000 ungrazed: 66 g C·m⁻², 2000 grazed: 0 g C·m⁻²).