Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Here Are Benchmark Results for Apple's New iPad Air With the M3 Chip

MacRumors

MacRumors

macrumors bot
Original poster
Apr 12, 2001
67,490
37,775


We are eagerly awaiting benchmarks for Apple's fastest-ever M3 Ultra chip in the Mac Studio, but we will have to settle with the first results for the new iPad Air for now.

M3-iPad-Air.jpg

The first Geekbench 6 results for the new iPad Air with the M3 chip are entirely unsurprising, but sometimes no news is good news. Given there were already Macs with the M3 chip, we know that it is around 20% faster than the M2 chip in terms of multi-core CPU performance, and the new iPad Air has similar gains as expected.

There are currently three Geekbench 6 results for the new iPad Air, and the device has achieved an average multi-core CPU score of 11,605 so far. The previous-generation iPad Air with the M2 chip has an average multi-core CPU score of 9,817, so the M3 chip is around 18% faster than the M2 chip in the iPad Air based on these early results.

This is all good confirmation to have, but as mentioned, it will be more interesting to find out how much faster the M3 Ultra chip is compared to the M4 Max chip. Those benchmarks should surface in the coming days, so stay tuned for that.

Article Link: Here Are Benchmark Results for Apple's New iPad Air With the M3 Chip
 
Article Link
https://www.macrumors.com/2025/03/06/ipad-air-m3-chip-benchmark-results/
MacRumors said:
This is all good confirmation to have, but as mentioned, it will be more interesting to find out how much faster the M3 Ultra chip is compared to the M4 Max chip. Those benchmarks should surface in the coming days, so stay tuned for that.
Click to expand...

I’m gonna go with M4 Max is faster in single core and M3 Ultra is faster for multi core.

A controversial take, I’m sure. :p
 
  • Haha
Reactions: brgjoe and G5isAlive
Plerf said:
I’m gonna go with M4 Max is faster in single core and M3 Ultra is faster for multi core.

A controversial take, I’m sure. :p
Click to expand...
It’s a tough choice! M4 Max is going to be considerably faster in single/low thread count workloads, but the M3 Ultra will smoke it for anything heavily multi-threaded (rendering, compiling etc.) or GPU bound (the 80-core SKU will get pretty close to 4080 Super in apps like Blender)

My heart says M3 Ultra 🙂
 
  • Like
  • Disagree
Reactions: thefrost and G5isAlive
The real question is how much faster the M3 iPad is compared to my current iPad with a A10X :p
And of course if I notice much of that extra speed in my daily use. My old iPad is actually fast enough for media consumption and internet browsing and editing small go pro footage such things. I only replace it due to the battery failing. Currently I consider trading it in or finding a way to replace the battery.
 
  • Wow
Reactions: aaronage
Honestly, I'm still happy with my A12Z iPad Pro 11 inch. I'd like a 13 inch, but fortunately (for my wallet) I haven't really seen anything, other than simply the bigger screen, to get me to upgrade to either a new Air or Pro. I haven't been blown away by Apple Intelligence on my phone, so that hasn't gotten to me.
 
roland.g said:
Let's be honest that the benchmarks on a MacBook Air for a general public consumption computing device really don't matter.
Click to expand...
Agree 100 percent

This is a non issue. Other features including battery life are way way more pertinent.

I’d even argue that a focus on getting something like a 20 hour battery life should be the focus in this category
 
andiwm2003 said:
The real question is how much faster the M3 iPad is compared to my current iPad with a A10X :p
And of course if I notice much of that extra speed in my daily use. My old iPad is actually fast enough for media consumption and internet browsing and editing small go pro footage such things. I only replace it due to the battery failing. Currently I consider trading it in or finding a way to replace the battery.
Click to expand...
I still have iPad mini 2 that I use at the gym to watch Netflix. Battery life is not very good. But not a bad device for what I am using it for.
 
Benchmark numbers don't mean much on iPad due to the OS. Anyway it should not be surprising to see the numbers considering that M3 has been out for a long time now.
 
  • Disagree
  • Like
Reactions: aaronage and mganu
PlayUltimate said:
I still have iPad mini 2 that I use at the gym to watch Netflix. Battery life is not very good. But not a bad device for what I am using it for.
Click to expand...
Yeah, the only reason I bought the iPad Air M3 was my addiction issues with apple stuff and the fact that in the last 3 hours my iPad battery went from 100% to 65% while the iPad was in sleep mode.
It's a waste and environmentally not good that we have to throw away (and trade in is throwing it away) a perfectly functional device that could be fixed with a $30 battery.
I looked up the iFixit Guide to replace the battery but that seems to be almost undoable for me. I would have to buy the battery and some tools for around 80 dollar and then I give myself a 50% chance to break something during the process. I have replaced batteries and screens and harddrives in Apple notebooks and iPhones. But the glued together iPad looks like a nightmare to open.
 
  • Love
Reactions: PlayUltimate
An M3 iPad is so utterly pointless that it has no reason for being. Until Apple gets rid of that junk iPad OS and run a real OS then it's no use upgrading a chip.
 
  • Disagree
Reactions: aaronage
Well, the M4 Max vs. M3 Ultra thing is not really hard to infer, at least in terms of Geekbench scores.

The M4 Max with 16 cores scores about 25600 points in multicore and 3900 in single core.

The M3 Max with 16 cores scores about 20900 in multicore and 3100 in single core. So, the M3 Ultra is unlikely to score more than double than that in multicore, which is 41800, even with perfect scaling.

In summary, M3 Ultra should be about 20% slower in single core and <63% faster in multicore, but only in heavily threaded apps; with 16 threads M4 Max could be faster, and probably until the thread count goes above 20 or so, the M4 Max will still be faster due to every core being 20% faster in the M4.

Same goes for the Metal Geekbench scores; M3 Max 16/40 scores about 155000 points, so M3 Ultra should get no more than 310000, and M4 Max 16/40 scores about 187000. Again a 65% gain. And in OpenCL, M3 Max 16/40 scores 94000 and M4 Max 16/40 about 115000. M3 Ultra should score less than 188000, so again a 63% gain.

These are, of course, best-case scenarios with perfect scaling, and assuming the M3 Ultra does NOT have a higher clock than the M3 Max.

All I can say is that this Mac Studio release makes me feel I made a good decision taking the Mini M4 Pro route a few months ago. :)
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back