Yes indeed: 5626 was manufactured to a higher accuracy standard than 5606:
5626 was "A" w/ Mean Daily Rate [MDR] of -6 to +9 secs/day
5606 was "C" w/ MDR of -15 to +25 secs/day.
And, as I said in Thian's post, an LM Special [5206 or 5216] was "B" w/ MDR of -10 to +15 secs/day.
At the high end, the 5646 was a GS that was manufactured to the "AA" standard of -3 to +5 secs/day [starting in 1969 & continuing until discontinuation of the mechanical GS manufacturing later in the '70s].
Also, beat rates for 5626 & 5646 were 28,800bph while 5606 was NOT hi-beat [21,600bph if I remember correctly].
Now, for jeweling - what Randall means is that some highly jeweled models used the "extra" jewels for either (or a combination of):
1. purely ornamental purposes - such as embedded into the rotor (or other areas) where NO movement wear would occur. Another use for jewels (cap jewels) is to keep dust out of a pivot (or lube in too?). If the jewel doesn't either minimize wear or keep out dirt, it likely is an "ornamental jewel". Things like the Walthams & Orients of the '60s or '70s with 64 & 100 jewel counts have lots of these.
2. marginally useful purposes, such as supporting the date ring as it turns [look @ some posts about the Seiko 62GS (6245 & 6246), 6218, and 6216 which all have 35 to 39 jewels]. Now, they may indeed serve a very minor purpose in making the movement "better" in operation, they are not critical to the long term viability of the movement.
So... yes indeed you should study the architecture of the movement to determine if the "extra" jewels (vs. a 17, 19, 21, or 23j movement) really add to the mechanical quality of the high jewel count movement [reduction in wear, better resistance to dust/dirt penetration...].
The Patron Saint of Seiko Collectors