I can confirm same for me - I have 2 displays, both the same resolution,. The left hand one is my laptop and the right hand one just an additional flat-screen. However, ordinarily the system chooses the laptop as the main monitor which I have changed so the additional display is my main as it is my preference due to that display being physically much larger. When I have the window on just the main display - maximized - it is fine other than when I toggle between maximized/restore it does flash momentarily onto the laptops display but as a slight overlap.
This overlap roughly matches an issue that existed prior to the latest update - a separate issue perhaps - when allowing resize of a window, the screen co-ordinates for the mouse pointer would have say about 20 pixels on the left hand side that have a negative value. Of course the right side always 20 pixels (maybe not exactly 20 pixels I don't recall) shorter than the reported resolution. Not only that, not matter what the virtual res was set at it would always start at 0 not the -20 pixels or so that it should start at. There was also no method to retrieve this offset that I could see. Whether this is a multi monitor issue or not I have not tested. In essence print commands show results 20 pixels in unless you go full-screen and there is no way of detecting this offset to make amendments to positions of text command in code or indeed sprite or pointer positions.
So far I have not confirmed the exact bug as described by OP - however it does do as described when I maximize the window onto the laptop display. It may not be obvious what the OP has described - those screenies threw me for a moment or two as I thought the right hand side of the screenies was a desktop background lol - this is because the overall screen shot matches my own display layout as I tend to use IDE on left also when compiling leaving my desktop background repeated on the right hand display...so that second screen shot looked initially to me as though you where running the window just on left monitor. I was also confused(and still am) what is meant by looking really zoomed in?)
EDIT: I want to be clear that although I state my desktop background is repeated it is just the wallpaper, the multi monitor setup is the desktop stretched over two monitors and not duplicate desktops, as I say just the wallpaper is duplicated
Having got past that I could see right away on my own system the screenies are accurate but one thing that did not come across in the description to me at least was the fact that the overstretch did not display anything other than the clear colour/clear screen command usage. A 3d cube and print() command only display on the monitor maximized and it looks dreadful. On my display it has a massive odd white border(see screeny). One thing that may be related is the fact you cannot drag the mouse pointer along the bottom or top edges of display from the larger screen to the smaller but can vice versa. Even though both displays are set 1920*1080 the mouse pointer gets stuck and remains on the larger display
which is OS behavior not agk, agk is just repeating this. There is a disparity between physical display sizes and maximum display res's for each - both my displays are maximum 1920*1080, also the laptop is 144Hz and the other display is 75Hz. There is no apparent correlation obvious to me regarding these frequencies as I have tried to set them both at 60 with no affect on the agk app results.
However that white border on the main window is NOT part of the window frame itself - the apparent "zoom" referred to by the OP is not just a zoom but an apparent separation of windows or perhaps an incorrect duplication internally?? Basically only the right hand side of my screeny shows a window title bar, and only by double clicking the right hand sides title bar can the window be restored, the max/restore/min/close buttons are missing, probably off screen to the right, the white border on the left display has zero window border/title bar access. Not only that but while repeatedly maximizing/restoring that white border on left display changes slightly - see second screeny. This is not the best part - the best part is that I have say chrome open behind the laptops display, the white section does not belong to agk - I can hover over buttons invisibly and the mouse pointer not only changes to match hovering over browser buttons, but gives me access to the browser, sometimes it does bring the browser to the front other times I can skip songs invisibly if I guess right location or change anything if I guess right from any web page! It was difficult to manage access of the thinner white sections but the bottom section works very easily in the manner I have described.
What I guess I am also saying here is that this isn't just a latest update issue. There is stuff going on that was related to windows prior to latest update and i feel that the results in the latest update are perhaps bad because of previous underlying issues - if you can't get the virtual res/pointer issue sorted I don't see how multi monitor setups would work, its only going to exaggerate any previous issues?
Win 7 Pro 64 bit SP1, AMD A4-5300 APU 3.4GHz, 8GB DDR3, NVidia GeForce GTX 750 1GB GDDR5, ASUS A55BM-E