Duffy's right.... sand is not all that a reliable or predictable medium for testing bullet penetration.
A whole lot of variables are involved; type of sand, moisture content, size of granules, type of bullet, caliber, etc, etc.
To answer your question, At one time I was curious enough to build a frame of 2"x2" lumber and partitioned it with 1/8" masonite partitions set at 12" intervals.
Using this setup, and common dirt with lots of "sandbox" sand from the open desert, we got the following results:
Softpoint magnum pistol bullets: much less than one foot.
Full-jacketed .30 caliber M1 Carbine, 9mm and .45 ACP stops in about one foot, maximum.
FMJ 5.56 M193 stopped in less than one foot.
FMJ 5.56 NATO stopped between one and two feet.
FMJ medium rifle calibers (220 g .30-40, .30-06, 7.62 NATO) gave several feet of penetration.
The 220 g .30 calibers gave the most penetration; upwards of 5 or 6 feet. (Shades of Karamojo Bell!)
Now, you know why the military universally considers two thicknesses of sandbags to be adequate protection against bullets and shell fragjments!
Oh, by the way, the US Military has conducted water penetration tests. ALL small arms bullets fired into water becomes harmless after only a few feet. Even the mighty .50 BMG bullets fall harmlessly after 10 feet of water.