We know the cost of SLS/Orion from public numbers and Government Budget Service.

We know the commercial prices for SpaceX launches.

Of course I have not done a full analysis of the cost of mating a service module and capsule in LEO. That would be far to much effort for a person that is not paid to do it.

But given that the difference is 3.5 billion, I don't think anybody could argue that this wouldn't be within that margin.

I would only hint at comparable programs cost. 3.5 billion is more then the complete Crew Dragon program, and we are talking about a dumb service module that consists of mostly things that already exists.

And that 3.5 billion $ is just the first launch, over the next 10 launches it would go up to about 15 billion $. Suggesting that it would cost more then 2 billion $ per launch to push a capsule from LEO to moon orbit is a bit absurd anyway you look at it. There would essentially be no new technology needed.

Some people have made series Artemis alternatives plans over the years, such as moon direct for example. But those of course change far more then my 'minimal' example.

But I of course have not paid millions of $ to do a complete study on an alternative architecture. That's what NASA should do, but they don't because they are not allowed to. And they aren't allow to because everybody knows it would be beyond embracing.

The only initial attempt to do a minimal study on just an alternative for Artemis 1 almost got the NASA Administrator fired.