It's my understanding that an accelerated IIgs is still meaningfully slower than a Mac for a handful of architectural reasons that you really can't get around unless you drop the 8-bit compatibility or re-architect the entire system so that 8-bit functionality is subservient to the 16-bit platform. (Similar to, say, how the Mac LC IIe LCPDS card worked.)
An Apple-built 16-bit multimedia platform re-architected so it was actually faster would have been real neat.
Idly, one thing I see a lot, and, actually experienced myself in school is that Apple IIgses were, in practice, used as faster/batteries-included IIes. We had a few IIgses at my school when I was a kid and they were used exclusively to run 8-bit edutainment software.
The problem any new computer platform introduced for or bought for k-12 education in the '70s, '80s, and '90s is that a school is typically a self-contained environment where everybody expects to be able to do anything they want on any machine, and, more importantly, where if you've got one or two rooms with 30 computers in them, you can't just sub 5 of those machines out for computers from a completely different platform because usually in K-12 the point of going to the computer lab is that everyone in the room is roughly speaking doing the same work.
In higher education you can get away with it more but higher education was more likely to have people using terminals to connect to big iron machines or to have students using their own machines. so you can get away with a mixed environment much more easily.
Anyway, at that point the question is... do you (Apple in the '80s) have a second machine up your sleeve for K-12 and that's probably why the IIgs ended up architected the way it was so it wasn't yet another attempt to move away from the II, but, more of an iteration -- with all the compromises that entailed.
By way of timelines:
- 1983 - Apple Lisa launches
- 1984 - Apple Mac launches
- 1986 - Apple IIgs launches
so I don't think "what if better IIgs" prevents Apple from doing anything with the 68000 unless you can get Apple a reliable 68516 at like 8-16MHz in like 1980. (That chip was announced/released in 1983 so the answer there is probably no.)
By way of interesting alt-history potential: I've always wondered whether or not a re-architected Apple IIgs would have been a better seller in a home environment where the potential for color and cdrom and faster performance plus integration with bigger/cheaper TVs as displays for some things could have been beneficial.
It would surely have been an interesting time to buy a computer. We had a discussion on here a few years ago about what you'd buy if you had $AMOUNT in 1986. I picked the Mac Plus because hindsight is 20/20, but the Apple IIe sold until like 1993 (outlasting the IIgs) and I think the IIc was around as well, so if you could get what you needed done on one of those you could get a solid productivity computer with a big ecosystem for less than what a IIgs or a Mac cost.
To that point: these platforms are really only as good as the software that runs on them and AppleWorks/HyperCard GS are "okay" -- they don't set the world on fire. They would've been terrible to use on a composite monitor and they're barely tolerable on the purpose-built RGB monitor. It feels like AppleWorks in particular is there to check a box relative to the Amiga for prospective Multimedia Authoring crowd. HyperCard, I haven't gotten into too much but it's got some nice ups over the Mac version that didn't get resolved on the Mac side of things until like 1997. Color is a big one, so I can see using HyperCard to build a CDROM title on the IIgs, but once you start thinking about going there - the costs pile up. (They would've on a Mac too, and, a lot of the story of the IIgs is that people wanted to do things that just weren't cheap/affordable practical yet.)