Right now on eBay
Marvel Super Heroes
Secret Wars #1 "Blue Galactus"
Newstand & Direct Error Variants
Marvel Comics, May 1984.Common
This is the the "Blue Galactus" variant of Marvel Super Heroes Secret Wars #1 (Newsstand and Direct Editions), it could equally be called the "Blue and White Galactus" variant or the "Galactus White Helmet" variant, although maybe the latter is another inappropriate double entendre.
This one is a curious error as it's not a simple color switch as seen on other error comics. As you can see Cyan (blue) is present and seems correct as is the key/black, also yellow is present in the mix, however, magenta is missing from where it should be but is present elsewhere (red is a magenta/yellow mix and there is plenty of red).
Above is a panel from the error variant, the corrected variant and a digital copy: the Thing should to be Orange (which is yellow with some magenta) but he appears to be red in the error variants.
Similarly, in the panel with Doom (above), all elements with any yellow appear red in the error variants.
What seems to have happened is that, on a few of the pages (first page and centerfold are prime examples), no magenta plates have been used and instead magenta is printed on another set of yellow plates hence all yellow areas are red and all pure magenta areas have no color at all so come out as white (the color of the paper).
It had been suggested that this wasn't an error at all but the colors were a result of the intense light from the Beyonder's power, however, a number of these panels (including Galactus taking off), are prior to the panels where Galactus and Doom are closing in on the Beyonder also lighter colors would be better for intense light (not red) so this all seems very unlikely (and, in any case, the error was spotted and corrected).
Corrected copies definitely exist but it is unknown how many of those were produced. The number of error copies is also unknown, however, both newstand and direct editions are affected and the error seems fairly common hence they were likely produced in numbers (with anecdotal evidence that some distribution areas got the error copies whilst others got the corrected copies).