While it's true that nobody can give a credibly very-accurate estimate, that doesn't mean such estimates are useless. We can look at similar engineering projects (similar in scale, timeline, and ideally similar in the regime) and make comparisons. For an international mission to Mars, two projects spring to mind: ITER, and the ISS. Both involved (or involve) a fair number of unknowns, both are decade+ international engineering efforts where lots of R&D was required up front, and both also include building large-scale structures and a consortium of partners.
The International Thermonuclear Experimental Reactor is the big fusion tokamok being built by France, Russia, the United States, Japan, and others, and the total cost is estimated to be roughly $50 billion dollars, with a total timeline of approximately 20 years, give or take.
The ISS (which needs no intro here!) cost roughly $150 billion in total, and was a roughly 12-year project, not counting ongoing use of it. It's a better analog for a Mars mission in some respects, because it involved designing and building a long-duration habitat for humans in space, in-orbit construction by docking together modules, multiple launches of boosters, and solving a lot of the same life-support problems that a manned Mars mission will have to solve. The ISS has 13 pressurized modules, which is (very likely) more than needed for a Mars mission, although they're only in LEO so they're much closer.
I didn't look up the $6 billion estimate you listed in the question, but I'm hazarding a guess that this figure is for a fly-by, and not a landing and ascent. Most of the estimates done by either NASA or outside groups have been higher. One of the most recent credible estimates is by an expert panel (see this National Geographic article from April 2014) which estimates $80 - 100 billion for a roughly 20-year program. Or, put another way, about 1/10th the total cost of the F-35 fighter program ;-)
When you read these cost breakdowns, the actual rocket launches are generally not a huge portion - even if we assumed a manned mission would require 4 x BFR or 4 x SLS Block II, at something like 2 billion/launch, and a complete dress-rehearsal unmanned, that's $16 billion or so. A lot of the money will be spent designing, building, and testing the craft that takes the astronauts there and back, the Mars ascent vehicle, and solving the "Mars landing problem" (slowing down heavy payloads enough to get them on the surface safely).
While every estimate should be taken with a big grain of salt - and prices vary wildly with different types of missions, number of people, length of stay on Mars, etc - there are entire disciplines in engineering around cost estimation, so it's not impossible to generate useful estimates.
Bottom line - somewhere in the
$50 billion - $100 billion range for a mission to put humans on the surface of Mars and bring them back, including R&D, and dependent on scope and ambition. I suspect that Elon Musk thinks he can do it for a lot less of course, lol.