The most logical order (assuming the limbs are roughly the same size and are analogous to the limbs of current animals) is to have the manipulation capable limbs near the mouth and the intermediate (capable of water propulsion) limbs at the back, leaving the leg analogues in the middle.
The reasons for this are simple: flippers are most efficiently used when ‘waved’ like the fluke of a whale’s tail. To do that well you need them to be free of obstruction, hence at the rear.
Manipulation is a very useful trait for accessing food but requires decent sensory input to be effective. Food is (I presume) ingested at the head end, so having the manipulators near the eyes/nose/mouth combo also makes sense. It’s much simpler to get a potential food item, manipulate it to liberate the food and then eat the food if your hands are near your face.
As evolution is strongly affected by energy efficiency (more energy for mating/killing if you don’t waste it elsewhere) it will tend to favour energy efficient solutions. Hence: fins at the back, hands at the front.
But since this is evolution we’re talking about you can posit damn near anything you want and justify it. If your creatures use their tails for propulsion and flippers for incredible directional control (like penguins) then hands/flippers/legs might make more sense. If they had an intermediate stage where the hands were used to feed their young in a pouch then legs/hands/flippers might make more sense. If they evolved over a long time scale then anything goes, really. Humans have some nerves that do loop-the-loops because our evolutionary path was long and complicated, and our anatomy is flat out wrong for bipedal living, yet here we are.
So really: whatever order suits you. Justify it later.
EDIT: just noticed the intermediated are also manipulation capable. My answer remains the same, but now you treat the creature more as a chimp than a humanoid. The legs are probably going to end up vestigial unless they’re essential for balance. The exact details very much depend on the complexity of the things they’re manipulating vs the amount and style of swimming they do though, so the second part of my answer is even more relevant.