What are the most and least sound gambits?

I thought 1...f5?? was the herpaderp gambit? It's a more fitting name. 

Sound = playable at the supercomputer level.  I guess strictly speaking that makes the king's gambit sound since they can walk the tightrope to the pawn down ending draw with ease.  In a king's gambit ending black's pawns are usually overextended due to an early ...g4 push, the king has easy access to the center (a trump in the Leningrad Dutch as well, even if the king ends up on h8 since g7-f6 can be used) and white has compensating piece activity and ironically king's gambit endings are easier to play than the middlegames though I still don't trust the KG in OTB human play... but maybe I should take it up again?  I don't know, sac a pawn, watch black overpress to exploit my kingside weakening, then use a slight initiative and slightly better king position in the ending could work out but black isn't guarunteed to overpress even if he does have a slight advantage with best play in the KG. 

An opening should be played with a mind towards the endgame, and KG endings typically lead to rook endings with plenty of pawns on both flanks. 

In my experiences, the soundness of gambits has a lot to do with how comfortable you are playing the gambit. There are many gambits that are, at the highest levels, considered "unsound" but at club levels you can play them if you're comfortable with them and understand the concepts. For instance, I know a guy who plays the Grob or the Spike opening, it's not a gambit but not popular at high levels of play but the guy knows & understand the reasoning behind the moves he makes and what his oppenent could counter with. He has destroyed many higher ranked oppenents b/c they underestimated the opening as a "coffee shop" attempt. If you're trying to really compete at high levels, study what the masters study. If you just want to have some fun try a gambit not very popular. The main thing is to understand why most unsound or unpopular gambits don't work long term, apply a conceptual understanding to them and figure out what you're comfortable with.  

Tarrasch said the Grob is a suicide move, and the point of a gambit is looking like a dashing player at the cost of losing the game.  The quote made me very sad since I was a big KG fan back when I read it. 

I played in a Grob thematic tournament on here, and with an ...h5 push black can develop the knight to f6 anyway. 

bongcloud > Grob lol

Smith morra gambit is sound, the problem is that black can decline and get a fine position. I think an other sound gambit is the two knights defence; 1. e4 e5 2. Nf3 Nc6 3. Bc4 Nf6! 4. Ng5!? white accepts the gambit, other posibilities are 4. d3, 4. Nc3 and 4. D4!?. After 4. Ng5 black plays, 4...d5 5.exd5 Na5! ( or Ne4!? With complicatipns..., Nxd5 is bad fx, 6. Nxf7!) 6. Bb5+ c6 7. dxc6 dxc6 and black usually plays ...h6 and ...e4 and gets absolutely fine compensation, activity!

It's definetly not for my repertoire, I rarely play gambits. Knowing which gambits are sound and which gambits are unsound might be useful in a real game where I have a tough choice to make (FYI I always take gambits if I don't know the lines) but I'm just asking out of general curiosity.

White's draw is far from easy, I agree... if you're a human facing equal opposition that is Otherwise machines just effortlessly walk the tightrope required.  King's Gambit really isn't that practical considering you could obtain the center and king activity later on in better openings. 

roman dzindzichashvili had a great video about playing against unfamiliar opennings and gambits.  I looked for it, but looks like it was taken down.  Though to paraphrase dramatically his arguement was if you don't know the theory just know taking the gambit will likely lead to a sharp position.  If you aren't ready/don't want the game to open up, refuse the gambit.  Of course he said a lot more stuff too but that was what I got out of it.

yes I can agree with that statement, but Never Call the KG unsound, as it is played by some very strong players, and has a history of proving a tough nut to crack, Fischer's Refutation is hardly that, It even makes a Double Muzio Style attack more powerful from what I've found against my cell phone, admitedly very weak. but now that I have an engine, I can analyse it more in depth.

here is the line I am talking about

I have beaten my phones most difficult setting (which normally beats me BTW )

1.e4 e5

2.f4 exf4

3.Nf3 d6

4.d4 g5

5. Bc4 g4

6. 0-0 gxf3

7.Qxf3 Qf6

8. e5 dxe5 

9.dxe5 Bc5+

10.Kh1 Qxe5

11. Bxf7 

For some reason The diagram thing is bugging out for me, I am currently using chrome, so it must be either my computer or my script, but you can find some short analysis of this on my most recent thread, and also planning on starting a thread to analyse this probably unsound attack

I agree with what you said.  he didn't say to universially decline a gambit if you are unfamilar with, but to recognize that if your opponent has offered a gambit, he likely prepared the line in detail. And you could get in trouble quickly if you don't have a plan and you let him take you down his pet line.

Someone said that the "Benko was probably the most sound gambit."  I'm pretty sure that the Benko has been refuted/close to refuted for some time now. 

http://www.chessgames.com/perl/chessgame?gid=1260002

I think this is the line, where black ends up down a pawn for pretty much nothing.  White can give it back anytime for whatever positional plus he so desires.  Thank God too, the Benko is annoying as shit if you don't know what you're doing!

That was over 10 years ago. Theory changes, ya know!  Anyway I seriously doubt the Benko is anywhere near refuted and I actually agree that it is probably the most sound gambit, as well as the Marshall in the Ruy. I wouldn't rule out the Gajewski in the Ruy either.

I think the most sound gambits are the Benko and Tal gambits, because they give black a lot of positional compensation.

I wouldn't include the Queen's gambit as a true gambit. Only beginners try and hold the pawn.

Although it is possible to hold the pawn for a long time, white will eventually have a much better position for it.

There are early black gambits where white trying to hold the pawn immediately gives black a huge advantage. Those are considered an opening trap to try and hold the pawn.

Drunken bushop

The reti gambit isn’t a gambit