Let's consider the the left-hand end of an open-ended tube oriented horizontally. Suppose that end were not an antinode. Then there are two possibilities:
1) The pressure a bit to the left, outside the tube, is even higher than the pressure at the end.
2) The pressure a bit to the right, farther into the tube, is higher than the pressure at the end.
#1 doesn't make a lot of sense; that would have the pressure outside of the tube, where it is free to expand up and down as well as sideways, higher than the pressure inside the tube where it can only expand sidewise.
#2 looks more plausible, but by the argument above the pressure just outside the tube cannot be higher than the pressure just inside the tube, so a higher-pressure region inside the tube will want to move to the left.
So #1 says that you can't have an antinode outside the tube; #2 says that the antinode nearest to the end of the tube but still inside the tube will want to move towards the end of the tube (antinodes deeper in the tube are being pressed from both sides so tend to stay put). Put them both together and you end up with the antinodes right at the end of the tube.