My gut feeling is that properties of air should be good enough (taking into account given mass fraction and 15% error). But I can be wrong.
Technically gasoline is a mixture (and not a well defined one - some of its properties are regulated, but not the exact composition). Its main components are hydrocarbons with 4-12 carbon atoms. See if you can find viscosity of the octane-air (or heptane-air, nonane-air) mixtures - while it will be still only an approximation, it should be a better one than using data for the pure air.