There’s nothing truly spooky about Apple’s new M3 chips — except, perhaps, for how scared they’ll make Intel, AMD and Qualcomm. During its “Scary Fast” Halloween Eve product event (at 8PM Eastern Apple, really?) the company officially debuted its new M3, M3 Pro and M3 Max chips. As featured in tonight’s Apple Evet, the M3 series will be featured in the revamped MacBook Pro 14-inch and 16-inch, as well as the 24-inch iMac, which never got an M2 refresh.
Apple is mainly positioning the M3 chips as major upgrades over its M1 hardware — if you bought an M2 system, you’re probably not itching for a replacement just yet. The M3’s GPU is the biggest leap forward, delivering new features like hardware-accelerated raytracing and mesh shading, which will enable more realistic lighting and better geometry handling. Apple claims the M3’s GPU is 1.8 times faster than the M2 and 2.5 times faster than the M1 in “Pro apps” — as usual, the company didn’t reveal its testing criteria. You can also expect better power efficiency, as the M3’s GPU can hit the M1’s performance level while using half the power.
Here’s how Apple’s new hardware breaks down: The plain M3 features an 8-core CPU (made up of four performance cores and four efficiency cores) and a 10-core GPU. Apple claims it’s up to 35 percent faster than the M1, and it can also support up to 24GB of unified RAM. The M3 Pro ups the ante with a 12-core CPU (six performance, six efficiency) and an 18-core GPU. It can squeeze in up to 36GB of memory, and Apple says that it’s single-threaded performance is up to 30 percent faster than the M1 Pro.
And then there’s the M3 Max, featuring a 16-core CPU (12 performance, four efficiency, a 40-core GPU and support for up to 128GB of RAM. Apple claims it’s up to 80 percent faster than the M1 Max. It also sports two ProRes engines to satisfy even the most demanding video professionals.
The M3 chips are also notable for being the first PC chips built on a 3 nanometer process, rather than the M1 and M2’s 5nm process. The increased transistor density helps with power efficiency, as well as providing better overall performance. According to Apple, the M3’s performance cores are 15 percent faster than the M2’s, while the efficiency cores are 30 percent faster.
Given that Apple just debuted the 3nm A17 Pro for the iPhone 15 and 15 Pro, it’s not too surprising that the M3 has been similarly shrunken down. In comparison, AMD debuted its 4nm Ryzen 7040 chip this year, and Intel plans to launch its Core Ultra Meteor Lake laptop chips in December, which is built on the “Intel 4” platform (using a 7nm process). The differences between architectures, some of which rely on newer tech like 3D stacking, makes it difficult to directly compare processing node sizes. But for now, Apple can lord its 3nm figure over the rest of the PC world.
When it comes to other upgrades, Apple says the M3’s Neural Engine, which handles AI tasks, is up to 60 percent faster than M1 chips. The M3 also sport a media engine with hardware acceleration for H.264, HEVC, ProRes (both standard and RAW). That engine also finally supports AV1 video decoding, which should make streaming AV1 content more power efficient.
Like most chip makers, it makes sense for Apple to follow up a major release like the M1 with a minor refresh like the M2. The M3 needs to prove itself to be the substantial upgrade over the M1 that Apple claims. And with the addition of ray tracing and better graphics, it may finally make Macs more enticing for developers and gamers alike. (Just in time for major titles like Death Stranding and Resident Evil Village to hit the App Store.)
Follow all of the news from Apple’s “Scary Fast” October event right here.