This paper is a continuation of earlier papers on the development of the SAGP process presented at Annual Technical Meetings of the Petroleum Society. SAGP improves the thermal efficiency of SAGD by adding non-condensible gas to the steam. Significant steam savings are achieved by lowering the average temperature in the reservoir and by reducing heat loss to the overburden. Rising gas fingers increase the pressure towards the top of the reservoir and displace oil downwards even though the temperature is below that of saturated steam. The gas hold-up in the reservoir equals the volume of the produced oil, with allowance for the effects of pressure, temperature, and partial pressure of the steam. The gas in the chamber comes from the combination of added gas, solution gas, and gas generated by chemical reactions, minus gas produced with the oil and the net gas driven to or coming from outside the depleted region by pressure difference. The gas hold-up and gas dissolved were estimated and it was found that, in general, more gas is required for higher pressures. Effects of layered sands on SAGD and SAGP performance are studied experimentally using a physical model. In SAGD, steam spreads below the low-permeability layers and oil cannot drain from above until the steam vapour can penetrate to replace it. In SAGP, gas fingers rise into the low-permeability layers and displace the oil downwards below steam temperature. Mechanisms for the enhancement of oil drainage from the low-permeability layers by gas fingers are discussed together with the experimental results. SAGP continues to show promising results and it is thought that results in the field will be better than in our experiments.