In GURPS the AR-15 and M16, from day one with the R601, have the standard Malfunction chance of 17.
What this means is you need to roll a critical failure when you're rolling to hit before you check to see if the weapon breaks in some way. See p. B407 for the rules.
Some guns, like a Glock 17, get a malfunction note of: "Very Reliable. Won't malfunction unless lack of maintenance lowers Malf."
There's five levels of malf in the rules.
Very Reliable 17+ twice (see High-Tech p. 79)
Reliable 17+ (default for TL6+)
Unreliable 16+ (default for TL5)
Very Unreliable 14+ (default for TL4)
Very Very Unreliable 12+ (default for TL3)
Spending the money for Fine (Reliable) (+25% cost) will move you up one step; Very Fine (Reliable) (+125% cost) moves two steps.
Historically, the Sten's malfunction rating of 16 can be traced to bad magazines, and the weapon's description mentions it.
In trying to find the introduction year for the M16 family's military issue magazines I started to wonder about their effect on the Malf. Especially looking at all the effort the Army (and the aftermarket) is taking to improve the magazine.
The Sten gets a one step reliability improvement if one finds good magazines... that means the design of the weapon is inherently at the default for TL6+, but there's a flawed part.
What if the AR family is inherently "Very Reliable" if you merely supply it with a better magazine (or just an inexpensive follower)? Full-30's mud test might be demonstrating that very thing.
If this is true, you can also say that you can make it Unreliable or Very Unreliable by supplying worse than USGI magazines.