Let me explain a bit better where is the contradiction. The high cost of source A from your example indicates that this source requires some expensive fuel. If there is an unlimited (in practice) amount of this fuel, in other words, it's not scarce, then this high price indicates that it's very hard to obtain (takes a lot of effort). This contradicts to the claim that source A has high EROEI, because if it has high EROEI, then it means that the fuel for it can be obtained very easily (with very little energy investment). (You could see the contradiction the best if you were to actually set the cost of energy in your world.)
The only way out of this contradiction is to assume that this "expensive fuel" is a very scarce (finite) resource, which cannot be readily produced from other things. For example, say it's platinum, and we would like to keep platinum to make other things out of it rather than burn it to make energy (and that's why it's expensive - people want to keep it). So that was my original interpretation of your example. In that case, source A is a better way to make energy than B, it's just you don't have enough fuel you would like to use for it, and so you can only use it in small volume. Just like with ROI - there can be excellent investments with very limited volume, and just ROI will not tell you anything about the volume.
As a side note - I really recommend to stop thinking about economics in terms of money and start thinking in terms of means of production (which is where EROEI comes in). And only then, when you see the full economic picture on the process level (how things get transformed in the economy), introduce the actual monetary accounting. This way, you can avoid some pitfalls (such as simultaneously claiming that price of energy is both very low and very high).
