Cursory testing on my 7-year-old Mac mini was OK in terms of speed, but the acid test (running Slack) still shows around 700MB of RAM in use by Firefox and worker processes—slightly less than Chrome, but almost twice as much as Safari (which was still leaner and faster), and there were some graphical glitches. Running it in Windows was OK, but it’s harder to make comparisons since my usage pattern there is so different.
Another thing that worried me was that Firefox seems to have a larger energy footprint and was still claiming a significant percentage of CPU cycles (5-10%) while out of focus and “idle”.
So, in a nutshell, good move, but still needs improvement—and I have so much personal investment in the Google ecosystem that switching away from Chrome’s identity integration and development tools will be very hard.