Stellarators have significant operational advantages over tokamaks as ignited steady-state reactors because stellarators have no dangerous disruptions and no need for continuous current &"ive or power recirculated to the plasma, both easing the first wall, blanket, and shield design; less severe constraints on the plasma parameters and profiles; and better access for maintenance. This study shows that a reactor based on the torsatron configuration (a stellarator variant) could also have up to double the mass utilization efficiency (MUE) and a significantly lower cost of electricity (COE) than a conventional tokamak reactor (ARIES-I) for a range of assumptions. Torsatron reactors can have much smaller coil systems than tokamak reactors because the coils are closer to the plasma and they have a smaller cross section (higher average current density because of the lower magnetic field). The reactor optimization approach and the costing and component models are those used in the current stage of the ARIES-I tokamak reactor study. Typical reactor parameters for a 1-GW(e) Compact Torsatron reactor example are major radius RO = 6.6-8.8 m, on-axis magnetic field B0 = 4.8-7.5 T, Bmax (on coils) = 16 T, MUE = 140-210 kW(e)/tonne, and COE (in constant 1990 dollars) = 67-79 mill/kW(e)h. The results are relatively sensitive to assumptions on the level of confinement improvement and the blanket thickness under the inboard half of the helical windings but relatively insensitive to other assumptions.