Like I said, the Ray Allen shot might be bigger. The Heat were certainly in worse shape at that moment than the Cavs were when Kyrie hit his shot, but even after his shot Miami still needed to get to overtime (the Spurs had 5 seconds left), and they needed to win overtime, and they needed to win game 7. If Miami had lost that game (or if they had lost game 7), I don't think it would even be a discussion over which shot is bigger (fair or not). It would go in that large bag of great plays by losing teams, possibly at the very top of the heap, but relegated there nontheless.
From an actual numbers perspective, Kyrie hitting that 3 was worth 27.8% WPA (Cleveland with the ball at that point was 58.1% expected to win, after the shot they were 85.9%). After Lebron missed his 3 and Bosh grabbed the rebound, the Heat were at 5.5% expected to win, and after Allen's 3, they were at 40.7% to win (35.1% WPA). Allen's shot was definitely [slightly] higher leverage just looking at the impact on that individual game, but considering Allen's shot was in game 6 and Kyrie's was in game 7, I'm still taking Kyrie's shot. A 40.7% chance to win game 6 (when down 3-2) I think comes out to around a 20% chance of winning the entire series, if a game 7 is a 50-50 proposition (might not be fair to the Heat considering the game was in Miami, but still). Compare that to Kyrie's 28% chance of winning the whole series, I'll take the latter.
*edit - since Jordan was mentioned earlier, his game-winner in 1998 was worth 34.1% WPA. That was a clinching game for them so it's higher leverage than Kyrie's, but I don't know how much to consider that if he misses that shot, the Bulls would get another chance in game 7.