This might become abundantly clear once the revised spell descriptions are released, but aallow me to ruminate.
Spells fall into rough categories as outlined by Bjorn, and generally become more powerful through two separate mechanisms; increased spell level or increased caster level.
As the caster level increases, the effect of the spells generally don't become more powerfull, but cover a larger area or have a longer duration. This increase typically has a corresponding increase in RP cost. The only saving is in the GB and Action component of preparing the spells. This allows higher level casters to be more 'efficient' with their actions and less dependent on courts for mass preparations, but the overall power level is sptill constrained by RP availability.
As spell levels increase it becomes less clear. Spells generally become more powerfull, but there is an increase in GB and/or RP cost. E.g. for spells that fall into the category 'zap unit on battlefield', is it intended that they become more effective per RP for the higher level spells? Or should the effect per RP remain relatively constant, and there again only being a slight efficiency gain in the number of actions required to acheive the desired effect?
Perhaps a way towards balancing the spells, other than lengthy playtesting, would be to consider a table of approximate effect per RP for various effects such as hits to unit, stat bonus to unit and prosperity bonus to province. Might be hard to set a fixed number to all things, and these are only some of the possible effects out there, but could be a good ballpark to work with for future spell creation as well.