The M2 MacBook Air is here and, like the M1, Apple has two options to choose from, one with an 8-core GPU and one with a 10-core GPU. We’ve been testing both models here at Macworld, the step-up model that starts at $1,499 (the model we tested and reviewed here has 1TB of storage and 16GB of RAM for $1,899), and the entry-level model with 256GB of storage, which costs $1,199.
That’s a pretty big difference in price, which led us to wonder what exactly you’re getting for the extra money. It turns out, you’re getting quite a lot.
With two fewer GPU cores, less RAM, and an SSD configuration that appears sub-optimal for read and write speeds, the cheapest MacBook Air faces several obstacles in trying to compete with its costlier sibling. In some cases we were shocked by how much worse it performed, but there were some pleasant surprises too.
Read on to find out how both the top-end and entry-level versions of the M2 Air got on in our speed and performance benchmarks. Where possible we’ve included equivalent scores for the 256GB version of the M1 MacBook Air released in 2020, and for the quad-core Intel i5 model that came out in the same year.
M2 MacBook Air: Raw processing power
We began by looking at processing power using the Geekbench 5 and Cinebench R23 CPU benchmarks. We didn’t expect a huge difference here, given that our 2022 Airs have the same processor with the same number of CPU cores, although the top-end model we tested did have twice as much RAM (16GB vs 8GB).
A promising start, with the entry-level Air tracking close to its more expensive sibling across all four tests (it was never more than 2.2 percent behind) and showing gains of up to 16.5 percent on 2020’s M1 Air. In fact the 256GB Air scored slightly higher than the 1TB model in Cinebench’s multi-core component, although this was by less than 1 percent and probably just an anomaly.
M2 MacBook Air: Real-world tasks
CPU benchmarks give you an idea of a machine’s speed on paper, but we want to know how these Macs will behave in the real world. We set them our usual battery of stabilizing, exporting, and encoding tasks in iMovie and HandBrake 1.5.
There was barely in difference in the M2 Airs’ speed when exporting a 4K file at High settings, but in every other test we saw significant loss of performance when using the cheaper model. Exporting at ProRes settings took 27.6 percent longer. Stabilizing an iMovie clip, meanwhile, took an alarming 43 percent longer; we actually had to wait longer than when using the M1 model from 2020 (which was also an entry-level unit and cost just $999).
In our HandBrake tests the 256GB model was slower than the 1TB version by 27.2 percent and 20.4 percent, respectively. In both cases it was closer in speed to the M1 model than to its own sibling.
M2 MacBook Air: Disk speeds
We test the read and write speeds of our review Macs using Blackmagic Disk Speed Test. We were particularly keen to see how the base Air got on in this test, following reports that its SSD is up to 50 percent slower on read speeds and 30 percent on write speeds. (The explanation, based on teardowns of the machine, is that Apple uses a single 256GB chip rather than two 128GB ones like last year.)
The 256GB Air was unable to disprove these dire prognostications. Read speeds were 47.9 percent slower than the 1TB model on average, while write speeds were a staggering 50.2 percent slower, far worse than expected. In both cases the entry-level Air is considerably slower than the 2020 model. You probably won’t notice the slowdown in normal day-to-day use, but for $1,199, we expect much stronger SSD performance.
M2 MacBook Air: Gaming performance
Last up, we looked at the two Airs’ performance running a pair of demanding games: Rise of the Tomb Raider, and Civilization VI. Both of these games include a benchmarking mode that allows you to measure framerates without the use of any additional software.
This is one test where Apple’s own publicly released specs for the two machines led us to expect a significant differential. Simply put, the 1TB Air we tested has a 10-core GPU, while the 256GB model has only 8 cores, so lower framerates are to be expected.
Test results were slightly confounding, however. Tomb Raider numbers were largely predictable, with the 8-core Air 43 percent behind at High settings and 26 percent behind at Medium. (In both cases framerates for the entry-level M2 were lower than for the M1 from 2020.) But the cheaper model punched well above its weight on Civilization, performing comparably to the 1TB MacBook Air with M2 on High settings and slightly better on Medium.
The top-end Air is undoubtedly a better gaming machine, but the loss of two GPU cores doesn’t appear to handicap the cheaper model as much as expected. Performance is likely to vary from game to game, however, and we advise caution if there is a specific graphically demanding title you plan to enjoy.
It’s scarcely surprising to discover that this year’s $1,199 MacBook Air performs less impressively than the higher-end model. What is more troubling is how much slower it is when it comes to read and write disk speeds (around 50 percent in each, according to our tests) and at real-world stabilizing, exporting and encoding tasks.
It’s always tempting to go for the cheapest configuration of a new Apple product in order to enjoy the new design and processor at the lowest possible outlay. As far as the M2 MacBook Air is concerned, however, we would advise against this, since testing suggests you’ll be getting a machine that in certain respects performs no better than a cheaper model from 2020, and in a few is actually worse. Admittedly, you’re getting a bigger and better display, MagSafe, and a new design, but the M2 performance boost just isn’t there.
If you do decide to buy the new Air–and we recommend paying extra for the upgraded configuration–make sure you find the lowest price by browsing our guide to the best MacBook Air deals. Or just pick up an M1 Air and save a couple hundred bucks.