Global Warming Potential (GWP) is an index used to measure the cumulative radiative forcing of a tonne of greenhouse house gas (GHG) relative to that of a "reference" gas (CO2). Under the Kyoto Protocol, GWP can be used as a fixed index to govern the trade-off between different GHGs in a multi-gas approach to GHGs abatement. The use of fixed GWPs has been criticized for not being very cost effective compared to the use of some flexible indices. To gain wider acceptance, however, a flexible index must also prove to be easy to use, and the economic gains from its adoption must be significant. In this paper, we develop a flexible index based on the concept of marginal rather than cumulative or average global warming potentials. These marginal global warming potentials (MGWPs) can be endogenously determined within a climate model given a particular climate objective based on radiative forcing level. The MGPWs are then linked to the marginal abatement costs of the GHGs, which are also endogenously determined within an economic model. When the two concepts are linked in this way, the result is a cost-effective way of achieving a particular climate change objective with multigas abatement. We show that the savings in costs when using this flexible MGPWs can be significant, and more importantly, they are not uniformly distributed across different regions.