GTmetrix - it will realistaclly show you how fast your website is, and focus is on that.
PageSpeed - focus is on making your website the best it can be. If they detect you could do something even a little better, like saving 10kb on image, they will punish you for that, no matter of speed load times.
I found this comment on reddit:
Pagespeed insights is probably the most misunderstood tool in use. The score you get from pagespeed insights has nothing to do with how fast your site loads.
Just look at the items they are using to determine your score. None of them are actual load times. The score is derived from all of the various recommended techniques to make a page fast. It is 100% possible to score very high and have a slow site or conversely score very low and have a very fast site.
Fun fact: its impossible to get a perfect score on Pagespeed Insights if your site uses Google Analytics. This is because you get docked points for not caching the JS used to load GA onto your site.
Looks at GTMetrix and the waterfall. That will help you identify issues with your site loading.
And from my experience, getting desktop results 90+, and mobile 70+ is pretty good on pagespeed.
Based on results that one of the fastest JS frameworks get, their website is optimized to the maximum by developers, NextJS results, everything that is even close to this is good
Also, is there a need to reduce the image size before uploading to Divhunt or the width and height option can reduce the image size without lowering the image quality?
No, lazy load might improve score but not much. You cant test it without lazy, and if you optimized your images correctly, you will still get amazing scores.
Divhunt doesnt do autooptimization of images right now.