Mind if we actually use some data now?
Let's keep it simple with wiki's list of estimates from various governments.The US works things out for plants and supplies coming on in 2016. This excludes any external incentives and most external non-regulated costs. The only nod to such externalities is a minor 3% increase in carbon intensive technologies.
I'd note that nuclear here is higher than coal because coal doesn't include external costs while nuclear does (it's heavily regulated plus healthy and safety risks are fairly direct, so easily estimated). The total costs of nuclear are projected as flatted (less variation) than any other source, which results in lower maximums and less variation.
Going to European estimates, far more external costs are rolled in. Coal is basically out of the window once EU regulations enter the picture. Solar is actually estimated as far cheaper in the UK, with a lot of regional variance, but still more costly than coal.
My problem with saying we need more nuclear is that we do need a good 5 years, minimum, to get nuclear plants online. Probably more. The massive slump in production means we have lost much of the trained labour we need to roll out new plants, meaning a lot of training needed. You don't want to ever cut any corners in the licensing procedure either. You can't really fast track new nuclear plants. We really needed to approve and start a new generation a decade ago. As it is we are losing an entire generation of plants with nothing much to replace them.
Affordable renewables don't really scale to cover the gap, while coal and gas are pretty stupid options even with capture schemes (I've only seen one that works and that requires building your plant on an old natural gas mine). Oil is a non starter. And we simply aren't going to cut energy consumption that much and really don't want to from a technological/social point of view.
Basically, no good answers.