It's hard to say from the outside because of all the optimism. They are targeting $250k per engine, and have thrown around numbers like $5-10m for the fuselage(?).

From what I understand, the engines are still >$1m each, but coming down as they ramp up production speed and have radically simplified the engines with V2. The body is probably also much higher than the target range, but bearing in mind that it's just stainless steel and relatively well known manufacturing techniques, and not aluminium going through complex and expensive manufacturing techniques, they're probably well under any other typical rocket. I'd assume even significantly less than a new Falcon 9, which is already the low end of the industry in terms of price.

Another thing I'd bear in mind is that many of these ships never got completed, on purpose. There have been explicit "pathfinder" builds, but even with the "real" ships most never flew and were pretty much about working through techniques, tools, materials, etc. These won't have needed a full fit-out and will have been cheaper.

Edit: updated engine target from 500k to 250k

SpaceX's Raptors cost about $250k each according to some leaked sources. Given there are 33 engines in the booster, you get 8,250,000 just in engine costs. I would say it's safe to assume around 30% of the cost of flight is in engines, so probably around another 5,775,000 for the rest of the craft bring it in at around $14M for just the booster hard costs. That's not accounting for the $100m crane they've built, labor, propellant, nor the Starship itself which has another 9 engines.

Keep in mind that the RS-25's cost $100m each. If it was atlas, the engines alone would cost $3.6B.

It costs six weeks of time, which is the only thing Elon cares about.

If the plan is to scrap the rocket no matter whether it succeeds or fails, does it cost anything? All of these seem like data gathering missions to me.