I picked up a Saab 9-3 recently for a daily now that I regularly transport a carseat. Definitely not to have a backup motor for the Goblin... but it does have the B207R and an F40. It's pretty fun and fits exactly what I was looking for in a beater.
In addition, at Gridlife, I heard from a couple folks asking about the Goblin, "you should put a 6 speed in it."
I suppose, now that I've driven an F23, F35, and F40, albeit with wildly different power setups, I can offer an opinion grounded in experience.
The F23 is the best.
The F35 was in my donor and in my Goblin with a pulley on the supercharger, so figure ~250whp. I drove it on the stock clutch and flywheel in the Cobalt and with a ZZP aluminum lightweight flywheel and Stage 3 Southbend clutch in the Goblin, and with a Cobalt shifter in both. It never showed any issues, no grinding or weird shift locks, but it never seemed eager to go into gear, and the RPM drop on the 1-2 shift took the fun out of putting around town. 4.05 ratio and G85 LSD.
The F23 has only had the ZZP flywheel + Southbend clutch and the optional DF shifter in the Goblin. I love it. The gear spread is perfect, the shifts are crisp and mechanical, and it's done great at Road Atlanta and Carolina Motorsports Park. 3.84 ratio and OBX LSD.
The F40 is still on its original clutch, pressure plate, and flywheel, with an OTTP short throw shifter. I don't fault the trans for the shifter being a bit more vague than the DF shifter + F23 combo in the Goblin even despite shorter and straighter shifter cables. The gearing is nice, but it doesn't make much of a difference to be in 4th or 5th between 35-50 mph or 5th or 6th above 45 mph. The worst part, though, is the heavy dual mass flywheel. I've not done any LSJ tuning for rev hang and I'm not sure if the B207R tuning encourages rev hang, but the revs just don't come down. I realize it's unfair to compare to ZZP's lightweight flywheel, but even compared to the original LSJ clutch/flywheel or other manuals I've driven (LS1, LS7, N55, C5 T56, GTI, BRZ, FiST, various Miatas), not only do they not come down, it feels harsh when the clutch and input shaft are more than a couple hundred RPMs off on shifts. An F40 dual mass flywheel reportedly weighs over 3x ZZP's lightweight F35 flywheel. Still, it's enjoyable for a stock daily 4 door car. This is with the 3.76 final drive.
Last but not least... a wannabe-nerd graph of relative accelerations. If you accept that RPM x gear ratio is a valid comparison of relative accelerations, then this should work... I think. Right? Take three hypothetical engines with equal and flat torque curves. If one has a longer gear ratio than the other two, and one is at a lower RPM than the other two, theoretically the longer gear ratios and the lower RPM motors should accelerate slower, at least until a gear shift. The below graph is shifting at 7000rpms.
One of the main takeaways is, the F23 and F40 are basically the same trans until 100mph and through gears 1-3. It's only above 100 that the F40 would pull harder than the F23 by basically splitting the F23's 4th gear in half. Based on this, I would've been a little faster on track by virtue of shorter gearing above 100mph, but is that worth it to me? Not on a mph/$ basis... I will add in the above chart, the percent acceleration comparisons are calculated without a shift if it's near the starting speed. So the 60-100 comparison is made starting in 3rd for all transmissions even though the F35 can do 70 in 2nd gear, and the 100-140 is calculated starting in 4th even though the F35 can do 109 in 3rd.
Additionally, in top gear and at 70mph, the F35 turns 2671rpm, the F40 turns 2517rpm, and the F23 turns 2496rpm.