This study shows how the assessment of emissions reductions from CO2 capture is critically dependent on the choice of multi-gas equivalency metric and climate impact time horizon. This has implications for time-sensitive mitigation policies, in particular when considering relative impact of short-lifetime gases. CO2, CH4 and N2O emissions from a coal-fired power plant in Brazil are used to estimate and compare the CO2-equivalent emissions based on standard practice global warming potentials GWP-100 with the less common GWP-50 and variable GWP for impact target years 2050 and 2100. Emission reductions appear lower for the variable metric, when the choice of target year is critical: 73% in 2100 and 60% in 2050. Reductions appear more favorable using a metric with a fixed time horizon, where the choice of time horizon is important: 77% for GWP-100 and 71% for GWP-50. Since CH4 emissions from mining have a larger contribution in the total emission of a plant with capture compared to one without, different perspectives on the impact of CH4 are analyzed. Use of variable GWP implies that CH4 emissions appear 39% greater in 2100 than with use of fixed GWP and 91% greater in 2050.