QGA Players which of withes third Moves do you consider most challenging?

QGA Players which of withes third Moves do you consider most challenging and why

3. e3

3. Nf3

3. e4

None. Because I know the answer to all of them. Hence why I'm a QGA player in the first place.

e2-e4 fails to impress because after e4 to e5, my knight just jumps to d5 - anti-Alapin style.

e6 is outright stupid - blocking your bishop. Let alone that it's not challenging.

Nf3... so I won't be able to play e5... I go and do other things, my normal move is e3 anyway, and not e4.

QGA is a choice of Anand, and it's practical. Highly underrated.

You want to know what would intimidate me: f3. - the Sämisch ('Zeimish'. King's Indian Defence, Sämisch Variation) pawn-structure:

After 2...dxc4, White's e4 and d4 pawns aren't scary - they're hanging, unless if f3 is played, That's serious - a g2-f3-e4-d4 pawn chain. But, since not One persons ever plays it, and since I don't think there is yet such theory, and people are cubes - doing only theory, especially d4 players - which are the more careful ones, I win with QGA.

f3-power in a d4 opening:

3:55

In summation, back to your question:

e4 is the most challenging, but only together with f3.

The irony is, that I learned the QGA from the 'no-f3' man:

i think 3.e3 followed by e5 is quiet good for white

i never heared about the f3 line in the e4 variation of the QGA

Just do the same thing against f3 as you would against e4.

I mean seriously, 3... e5 and what does White do? 4. d5 just looks abysmal, 4. dxe5 and White doesn't get to castle, Black gets an open file, and Black has some easy targets, and don't forget Black's still up a pawn.

If I were a QGA main, I'd be most alert against 3. Nf3 players, as you're pretty much just accepting a Classical Mainline or Gunsberg at that point, and I'm not a fan of either of those positions as Black, but I AM a big fan of them as White.

I'm currently fiddling around with the 4... dxc4 Slav, and as far as I can tell, it's just a better QGA, so why not give it a shot?

I feel like QGA is impractical because there's a lot of lines which avoid it

When would that ever be an issue?
3. f3 and 3. e4 (the only logical thing to play f3 with) both are countered by 3... e5, you can't play 3. Nf3 because you're going to put a pawn there, and barring 3. Nc3, any other third move loses almost immediately.

So your two options are:
Line a) 3. Nc3 a6 4. f3 b5 5. e4 Bb7 6. a4 c6 and
Line b) 3. Nc3 Nf6 4. f3 c6 5. e4 b5 6. a4 b4

Neither of the above are good. Sure, you may have two pawns in the center, but Black has much better development, is up a pawn, and has a rather strong queenside attack and even potentially action on the kingside due to just how abysmally weak f3 is. There's one line in line a where Black plays g5 unprotected in plain view of two Black pieces and proceeds to play g4 gxf3 Rg8 Rg2 Bg7. No human would ever see that, nor would any human reach the position to even play g5 in the first place, but the f3 QGA is a laughable concept. White's LSB and KN are terrible and even the QN and DSB aren't great, White's center isn't even that strong with pawns on c7/c6, e7/e6, and f6, Black is strong on both sides of the board. White is so close to simply losing on the spot at any time due to the amount of pressure coming from Black.

all three of those responses are good👍.

But to answer the question, I find 3.Nf3 to be the most solid to play against.

"QGA Players which of withes third Moves do you consider most challenging?"

++ Former USSR Champion and second of Kasparov, GM Iosif Dorfman considered 3 e4 most challenging.

Back when I tried to play the QGA, the 4.Nf3 line with I think it was 7.Bb3 that was a problem for me. I've moved on to better openings.

It definitely wouldn't be e3, since at the very worst that can be transposed by black into the main Nf3 line, and otherwise black has some interesting alternatives. 
But if I had to choose between Nf3 and e4 ... I'd say e4 is more challenging. Mainly because there's alot of room in the positions for white to explore and come up with moves, and the positions can get chaotic... black doesn't have a way to respond that leads to a stable, predictable game like with Nf3. Not to say the Nf3 positions aren't interesting, it's just that black consistently gets that position on the board and can study it in depth, it's not easy to take the game into unexplored territory.
That said, e4 is probably the hardest for white to play as well because black has many different possible responses to e4...

I agree with your e3 assessment, but I disagree with your e4 assessment.

Black essentially has one way to get chances against 3. e4, and it’s with 3… e5. Anything else is very good for White, as you end up equal on material but White is up on development, space, and center control.

I always wondered what do QGA players play against c4 and Nf3?

There are 5 moves played at master level by black after 3. e4 here. Regardless of your feelings about the positions that arise, white has to know how to respond to all of them ... so this puts a greater burden on white, and hence this is harder for white to play. I don't care what the engine says, whether you control the center on move 4 or not - any complex, novel position that you're intimately familiar with and your opponent isn't... is a significant advantage to you. Which is why openings like the Pirc exist. In contrast, there are openings that follow classical chess principles but which put almost no burden on the opponent to find correct moves since every move just... follows the obvious chess principle. If I'm going to bother studying an opening I want it to put pressure on the opponent to find correct moves. Because at the end of the day... chess is a draw when everyone plays perfectly. Even in less than ideal openings. 
Infact at master level 3... e5 doesn't have a great winrate for black compared with 3... Nf6, 3... b5, or 3... Nc6, probably for precisely this reason - 3... e5 is the most common move. But at my elo, for now, I prefer 3... e5 ... but as elo increases that will probably change.

"Most challenging" (and this does not mean "best", actually e3, e4 and 3.Nf3 are all equally good) is certainly enough 3.e4, which leads to sharp play.

3...Nf6 and 3...Nc6 are also good, and the currently 3...b5 is very trendy at GM level.