I'm in the market for a MacBook, primarily for content creation and full-stack development using tools like Next.js, FastAPI, PyTorch, YOLO, Docker, and Redis. My office provides a separate computer, so this one is for personal use only. I'm debating between getting a MacBook Air with an M5 chip and 24GB of RAM or a MacBook Pro with an M5 chip and 16GB of RAM. I'm concerned about the Air's potential heat issues and throttling during heavy use, as I've heard mixed reviews about it. Any advice on which model would be better for my needs?
7 Answers
I've been using the Air with an M5 and 24GB and it handles everything just fine. Just a heads up though, the USB-C ports are only on one side, so that’s something to consider!
Definitely choose the Air with 24GB of RAM! I have the Pro with an M3 chip and 18GB, and it can get sluggish with 10GB of swap used. The extra RAM really helps with multitasking.
Go for the Air M5 with 24GB. For your use case, more RAM will be more beneficial than having a fan. The throttling issue is exaggerated unless you're pushing it for hours nonstop.
Seriously, go with the Air M5 and 24GB. Throttling is really a non-issue for most dev work, and having the extra RAM is key when running Docker and PyTorch simultaneously.
I switched from a 2019 Pro to an Air with max specs, and it's been fantastic. The difference in weight and size is significant, and my company covered the costs, so I was able to save some cash too!
I'd lean towards the Pro. You might want to explore refurbished options, though. Models with M3 or M4 chips are still solid and usually cheaper too.
If you can afford the Pro, just go for it. The extra power and cooling can make a noticeable difference, especially if you plan to dive deep into development.
Honestly, I think you'd be better off with the Air that has more RAM. A well-equipped Air can handle most development tasks quite nicely without overheating.

That makes sense, but keep in mind that the 16GB Pro actually costs more than the 24GB Air. It’s easy to pay extra for cooling but giving up RAM might not be wise for your workload.