Mate in 2, white to play

here the correct one

I'd like to know why she says in the video below (to Magnus Carlsen) why this is mate in 2, the engine says mate in 3 (3 options)

https://www.youtube.com/watch?v=D-x3Wzy6pVk

Your diagram is wrong. Remove Rc5.

it is correct as it is, I double checked it

yes you are right I'll replace it

I didn't say you did anything wrong but the diagram is wrong. You probably copied it from where it was wrong in the first place. You can test with StockFish that the puzzle has several solutions and that can never be right.

Btw, if you didn't know it by "checking" you might think that 3 white rooks is a bit crowded for a puzzle. Serious puzzles commonly stick with the material in your initial chess box.

that is my issue: M3 it wrong (although this works as well [obviously])

Your revised diagram is OK. When I ask the chess.com engine, it says "M2". No clue why your engine would say "M3".

indeed strange due to the fact we have the same engine as you do. 

I got the answer: if you change the engine to Komodo, it says M2

Perhaps you let black start instead of white. Then it is mate in 3!

Not for me. SF already gives me the correct analysis and I need no Komodo to confirm it. Note: you should only look at the first line. That is the best move and the only correct one here.

1. Re3+ Kd4 2. Ra3+ Kc4 3. b5#

I couldn't spot nate in two

Try the Ruy Lopez or the Queens Gambit

Solution: MATE IN 2 

 1. Qa8 Rc4 2. Nd6#

Is there someone who can explain to me, why SF can't detect M2 and only M3? Ariskotle: I posted my proof above (M3 for SF (Stockfish). 
Thank you.

I got the solution by myself: you have to activate the could analysis to get M2:

If you leave it as SF11 and change the depth to "unlimited" it also comes up with M2 after less than a second. (That's immediate if I run SF11 in Tarrasch on my desktop.)

I also get the M3s with SF11 without setting "unlimited".