what’s leaked? All I’ve seen is complete conjecture.
https://www.macrumors.com/2021/05/18/bloomberg-mac-pro-32-high-performance-cores/
From Bloomberg article.
"...Redesigned MacBook Pros are expected to debut as soon as early this summer, said the people, who requested anonymity to discuss an internal matter, followed by a revamped MacBook Air, a new low-end MacBook Pro and an all-new Mac Pro workstation. .."
When someone with an Apple badge talks to a reporter and says the upcoming systems have "blah blah blah" then that is a leak (from Apple).
back in December.
"... Apple is also reportedly testing a chip design with as many as 32 high-performance cores for higher-end desktop computers planned for later in 2021, as well as a
new half-sized Mac Pro planned to launch by 2022. . ..."
Apple is working on a series of new custom Apple silicon processors to power upgraded versions of the MacBook Pro, new iMacs, and a new Mac Pro for...
www.macrumors.com
And earlier report
"... Apple's first Mac processors will have 12 cores, including eight high-performance cores and at least four energy-efficient cores, according to the report. Apple is said to be exploring Mac processors with more than 12 cores for further in the future, with the company already designing a second generation of Mac processors based on the A15 chip. ..."
In line with a timeframe shared by analyst Ming-Chi Kuo last month, Bloomberg today reports that Apple is planning to release at least one Mac with...
www.macrumors.com
Technically the above doesn't fit the recent updates of a 8p-2e-16g ( and 8p-2e-32g ) building block that the recent report talks about. But if just look to the "Jade 2C" and "Jade 4C" ( where 'c' could be chiplet)
you get 16p-4e-64 with that larger baseline "doubled up". ( so now at minimum of 4e )
The oldest report was suggestive that they were going to use a "8p-4e" building block. It wouldn't be too surprising if they either walked that backwards to get to either
i. smaller (more affordable ) pieces
or
ii. that leaving some cores out ( flipped off for slightly better yields ( A12X -> A12Z over time) and/or bandwidth issues with the large number of hungry GPU cores attached.
So about over a year with Apple person saying about the same thing three times in a round.
As to whether Apple is using chiplets or monolithic dies or some method between the two perhaps can label that conjecture. There are multiple ways to skin the cat here (GPU cores on their own chiplet . GPU cores spread out over multiple dies , GPU cores on one huge monolithic die. ) TSMC has 2-3 multi-chip-module techs Apple can use to cobble together different SoCs from the same baseline die or just do various monster big ones. ( Previously Apple only went to about 120mm with iPad Pro SoCs. the M1 lands in that zone and not surprisingly also landed in a iPad Pro. )
But is the MBP 16" likely coupled to the iMac and "half sized" Mac Pro? If they have related code names Jade 1C , Jade 2C , and Jade 4C ... then the names are certainly coupled. ( some M1-adjective bigger SoCs. )
That the new MBA , MBP 13" , (probably entry Mini) ,etc. hare a different code name makes them pretty likely to be 'M2'
The RAM being soldered on is a bit of an inference. But how many mid-large GPUs have you seen wither DIMM assigned RAM? As the core count grows much larger keeping the cores feed with data becomes an even more complicated balancing act. Arbitary DIMMs would extremely likely make that even more difference. That's more conjecture on why arbitrary user RAM would actually work more smoothly.