From a PM answer of James Hooker, about my question about buffer gas stating of the modern T5 HO/HE on his site , It turned out that the gas of the modern T5s, is just a pure argon, just like the classical T5 4-13W, the T12s and T8 of 15W, 20W, 30W and 50W (So they have the same starting voltage as the old fluorescent lamps), and that the reason for their higher efficiency over short T5s, is because the pressure of the mercury is kept nearer optimum levels, due to a long mount at one end (The reason why in long T5s, one electrode is longer than the other one, causes larger blackening in that area, when the tube is old.
Also James said me that the red and green components of the triphosphors are usually better than the triphosphors of the T8 (Probably the reason why current triphosphors T8s are barely brighter than halophosphors T8s).
There is also less self absorption of the mercury 253.7nm UVC line, because of the narrower tube diameter.
If this is the case, so why the optimal temperature of the modern T5s is 35degC (95degF), and not room temperature? (In room temperature, they are less efficient than the T8s).

And of course, the reason why they are limited to electronic ballast, is because manufacturers refused to produce magnetic ballasts for them, because of the time of their invention (14W and 21W HE T5s can be operated successfully on a preheat/switchstart magnetic ballast for 13W classic T5s).

Edit: And why the PLs, CFLs and 2D lamps have argon-krypton mixture like the T8s and not pure argon like the modern T5s?

What I observed, even on T8 the efficacy difference between halo- and tri-phosphors is about 20..30%, what is not as low figure.

The increased rated efficacy of T5 compare to T8 rating come in big part from the fact, then the T8 are rated for 50/60Hz supply, where the T5 for HF supply (as they are not rated for mains frequency), so in fact the rated fifgures are not comparable.
If we take a F36T8 as an example:
It is usually rated at about 3300lm on power input of 36W at 50Hz, so ~90lm/W.
The same lamp on HF for the same 3300lm suffice with about 32W, what mean ~103lm/W, what is nearly the efficacy of T5 with similar output.

But in real life the T5 benefit (or better say do not loose) mainly from the more optimum temperature:
In real fixtures it is nearly impossible to keep the temperature of the air in the fixture to be at 25degC, so the T8 normally run hotter then optimal, or need open, well ventilated fixtures, whose then tend to collect dust.

The 35degC of the surrounding air is easier to achieve and maintain with enclosed fixtures, at least indoors (but where the majority of fluorescent lighting is used).
So the T5 fixtures run way closer to the tube optimum temperature then T8 ones, yielding higher efficacy of T5 compare to T8 fixtures.