This paper investigates the effect of the gravity framing system on the overstrength and collapse risk of steel frame buildings with perimeter special moment frames (SMFs) designed in North America. A nonlinear analytical model that simulates the pinched hysteretic response of typical shear tab connections is calibrated with past experimental data. The proposed modeling approach is implemented into nonlinear analytical models of archetype steel buildings with different heights. It is found that when the gravity framing is considered as part of the analytical model, the overall base shear strength of steel frame buildings with perimeter SMFs could be 50% larger than that of the bare SMFs. This is attributed to the gravity framing as well as the composite action provided by the concrete slab. The same analytical models (i) achieve a static overstrength factor, Ωs larger than 3.0 and (ii) pass the collapse risk evaluation criteria by FEMA P695 regardless of the assigned total system uncertainty. However, when more precise collapse metrics are considered for collapse risk assessment of steel frame buildings with perimeter SMFs, a tolerable probability of collapse is only achieved in a return period of 50years when the perimeter SMFs of mid-rise steel buildings are designed with a strong-column/weak-beam ratio larger than 1.5. The concept of the dynamic overstrength, Ωd is introduced that captures the inelastic force redistribution due to dynamic loading. Steel frame buildings with perimeter SMFs achieve a Ωd>3 regardless if the gravity framing is considered as part of the nonlinear analytical model representation. © 2014 John Wiley & Sons, Ltd.