M5 Max feels like cheating

The M5 Max chip is probably the most impressive piece of hardware I've tested in a really long time. This thing honestly feels like a cheat code, but you also shouldn't buy it. The performance that Apple has been able to squeeze out of these 18 CPU and 40 GPU cores is truly mind-boggling. Apple is seriously crushing it right now. But one of the biggest things that they're going to crush in the case of the M5 Max is your wallet. A 14-inch MacBook Pro with the 20 core M5 Pro, 24 GB of RAM, and one terabyte of storage is going to set you back $2,400. But if you want to upgrade to the base version of the M5 Max, you're going to need to tack on $1,200 because you're not just buying the extra GPU cores, you're also getting 36 GB of RAM and 2 TB of storage. But if you want to upgrade to the full 40 core M5 Max, you're going to need another $500 because Apple also requires you to upgrade your RAM to 48 GB. So if we equip an M5 Pro with 48 GB and 2 TB, that's going to cost $3,200. So that means that the full 40 core Max chip cost $900 just on its own. Now, you're getting more than just 20 extra GPU cores. Memory bandwidth doubles from 307 to 614 gigabytes per second. You also get two video encode engines and two ProRes encode and decode engines as opposed to one. But that is a very steep price to pay. So the question is, is it worth it? Well, to find out, I tested nearly every single Mac that I own. And on top of that, I'm using a new service that I launched last week called MeanBench. This is a platform that I'm working on to effectively crowdsource benchmark data. I've been gathering information from you guys, testing a bunch of benchmarks yourself at home, and I've synthesized that into what is probably the most comprehensive set of benchmarks you've ever seen for these MacBooks. So, let's get to the bottom of the M5 Max chip right after a word from today's video sponsor. Does your day-to-day carry look something like this? If so, then today's video sponsor, Broy, is here to help. Broyy's classic day pack is designed to be able to swallow everything you throw at it while also looking like a classic elegant backpack. On the outside, you've got two long pockets for bottles along with a handy front pouch. Inside is enough room to carry multiple books as well as a laptop and tablet and a secret compartment for your small valuables. Speaking of smaller valuables, I never know where to store my passport or driver's license when I travel. But thankfully, Broy has a solution for that as well. Their excellent travel folio has space for two passports, eight cards, hidden pouches for cash, and it even comes with a pen. And the best part about both of these, and indeed all of Broyy's products, is that they focus heavily on materials and build quality. Most of their fabrics are 100% recycled, and they source their leathers from leather working group Tanneries. I highly recommend Broy stuff. In fact, I've been using them for years. They probably didn't even need to sponsor this video for me to say that. But if you want to save 10%, you can use my link in the description below. A huge thanks to Broy for sponsoring this video. And now, let's get back to it. All right. Well, I'm not going to waste any of your time. Let's jump straight into the performance figures. First of all, in Cinebench 2026 on the multi-core score, I was actually a little surprised here. The M5 Max actually scores a little bit worse than the M5 Pro with the same exact CPU, but most likely because of the increased thermal envelope from this huge processor, it wasn't able to run that CPU quite as hard. That said, however, the jump this year in the Pro and Max chips compared to last year is unbelievable. This is an insane year-over-year update, and we see the same results carrying over in Blender Classroom on the CPU test. And compared to last year's M4 Max, we're only gaining about 6 seconds. The M5 Pro was definitely the bigger gain with over a minute difference between the M4. But of course, the whole reason to buy the M5 Max, is for the GPU. So, how does it do? Well, in Cinebench 2026 on the GPU score, which I don't have on my bench, so these are just my local devices, the M5 Max is unbelievable. This is by far the most powerful GPU that I have tested. That's 50% faster than the M4 Max and more like 60% faster than the M3 Max. Both of which have 40 core GPUs just like the M5 Max. So these are just per core improvements. That is freaking crazy. And if we bring in a bunch of other devices over in Blender classroom on the GPU render, this thing is unbelievable. 9 seconds. I mean, look at the M2 Ultra with 76 GPU cores, 24 seconds compared to 9. I even fired up the render on my gaming PC with an RTX 3090. And even with Nvidia's crazy optimization and the 350W TDP of that graphics card, the M5 Max in a laptop still beat it. What is the Apple silicon team smoking? And can I have some? Next up, we've got Shadow of the Tomb Raider at 1440p on the high settings preset. And I love testing with this game because it predates Apple Silicon, so it's not perfectly optimized and it doesn't take advantage of ray tracing. So, just raw GPU power here. And you can see the M5 Max at 102 FPS is still improving pretty significantly over the M4 Max with the same number of cores. But the M2 Ultra, it just has so many cores to throw at the problem that in this case, it outperforms the M5 Max. But in Death Stranding, which is more optimized and has launched with Apple silicon support explicitly, well, this is able to take much better advantage of these GPU cores. In 4K on the high preset, the average FPS is over 100. That's perfectly tied with the M2 Ultra despite having half the cores. And even the 80 core M3 Ultra is barely doing better. Also note the difference between the M4 Max and the M5 Max. The more optimized the program is, the better the M5 Max is going to do. So, if you throw an optimized game that also has ray tracing at this thing, it's going to blow everybody else out of the water. In Cyberpunk 2077 on the 4K ray tracing ultra preset, which it's worth pointing out does use upscaling, so this isn't running at native 4K resolution, but nonetheless, the M5 Max is once again blowing everybody else out of the water. were nearly double the FPS of the M4 Max and almost exactly double the M3 Max again with the same number of cores. And I know I keep saying that, but it's nuts that in two generations, Apple has doubled the ray tracing performance with the same cores. And even if you take ray tracing out of the equation, going back to 4K medium settings, well, here we find that the M5 Max again is clobbering everybody else. the M2 Ultra just doesn't have the optimization of these cores to keep up. Now, again, to underscore how powerful this chip is, I ran the same test on the RTX3090 and it scored 52 FPS. So, that is a little bit better. And running the test again on the medium preset, the M5 Max is 98 FPS, not again too far off the 123 that we get on the RTX3090. And again, this is a laptop chip that is in Mac OS, not nearly as optimized as a 350 watt Nvidia GPU running Windows 11. So, this is pretty crazy. And I think that's the most interesting thing when you're looking at Apple silicon performance because the application really does matter. And of course, one of the key areas to apply that is in AI. Apple stressed how powerful these machines are for AI with the new neural accelerators on each GPU core. And sure enough, when I downloaded Olama and ran a couple of tests with the Quen 330B model, looking at the M5 Pro and Max side by side, you can visually see how much faster the Max chip is running. Now, at the end of the day, this is not a huge difference. We're talking about a matter of seconds here. But I ran the same prompt on a bunch of different machines a couple different times and averaged how long the model was thinking for, and that revealed a pretty interesting result. The M5 Max was the clear winner, even beating out the 76 cores in the M2 Ultra. But I think what was most interesting, frankly, was how well the M5 Pro did. Cuz my 16 core Bend M4 Pro in my Mac Mini, it took ages compared to these machines. I mean, the new Prochip is up there with the Maxes and the Ultras. As fast as the M5 Max is, it was 2.8 seconds faster on average than the M5 Pro. Is that worth $900? I mean, maybe if you really scale it up and do a ton of AI workflows that that would add up, but I I don't know, man. 900 bucks is a lot for that. And moving over into video editing, well, it's more of the same, frankly. We we talked about this yesterday with the M5 Pro, but Final Cut Pro has been so well optimized thanks to the media encoders that you get on the Pro and the Max chips that they're all pretty much the same. And nothing changed this time around either. The M5 Max rendered out my benchmark timeline in 46 seconds, exactly the same as the M1 Max. And that brings me to something that's always kind of confused me about the Pro and the Max chips. We know that the Max chip has two encode engines for video and for ProRes compared to one on the Pro chip. So, why is it not any faster in the render? And you might think, well, maybe the export is where you see the gains. And well, kind of. I mean, we're definitely seeing some improvement. The M5 Max shaved 9 seconds off compared to the M4, but the M2 Ultra beats everything. It's at 46 seconds. But frankly, that just gives me more questions because we didn't really see that many gains going from 1 to two. So why do we all of a sudden see huge gains going from 2 to 4? Maybe it's cuz it's two distinct chips that are able to operate in parallel. Whatever it may be, we saw the same results in Da Vinci Resolve with a minute and 9 seconds. Yes, you're saving some time compared to the M4 Max, but even as crazy as the Max chip is, you just can't compete with having two processors. So, the M2 Ultra takes the cake here. So, the M5 Max is a freaking beast. I mean, this thing is just crazy. But you might be wondering at this point, having all of this performance in the exact same chassis has to have some caveats. Well, to find out, I ran Cinebench, both the CPU and GPU tests backtoback a couple of times so that I could get an idea of the thermals and the battery impact of the Max chip. And first of all, when running the CPU test, you can definitely see the Max chip is actually running a little bit hotter. This probably explains why we saw a slightly lower score compared to the ProChip. But after running Cinebench backto-back for 20 minutes on both of these machines, the battery was about the same. 87% on the Max and 86% on the Pro. So, let's run the GPU test now and see how that does. I'm seeing hot spots at a little bit over 47° C compared to a little over 43° C on the M5 Pro. But running the test back to back three times absolutely shredded the battery on my M5 Max. While the Pro was sitting at 54%, the Max got down to 30. Now look, I'm neither surprised nor complaining about the impact to battery life. I mean, it's kind of what you would expect when you just pack extra GPU into the same chassis and battery. So, honestly, I think it's pretty impressive that Apple is able to run this chip without really impacting the thermals all that much compared to the Pro. But if we zoom out and we look at both of these chips side by side, I think the big question is, did the M5 Max justify itself? It's a $900 upgrade just to get those extra GPU cores and the memory bandwidth and those extra encoding engines, which didn't even really do all that much. So, to be honest, I think that my conclusion here is similar to last year with the M4 Pro and Max. And again, to be clear, you should not buy either one of these laptops because we're going to get new models in the fall. But I suspect with the release of that new generation, the same recommendation is going to be true. And that is that if you want a Mac chip, what you really should be getting is a Mac Studio. The Mac Studio is a really, really good value for the amount of performance that you can get for $2,000. And even if Apple increases the price by like two, three, $400 to put the Max chip in, which would also come with commensurate increases in RAM and storage, I think that that is just the way to go. And when you look at it from that perspective, I think the big lesson is forget about the Ultra chip. The Max chip in the Mac Studio is probably the best value overall cuz the Ultra chip lags behind. Right now, if you buy a Mac Studio, you get an M3 Ultra. That's two generations behind now. And as we saw in Death Stranding, you get an extra what, 10 FPS for double the cores in a Mac that costs $5,500. Yeah, dog. I'm just going to wait like 2 3 months, get the M5 Max for 2,400 bucks in a Mac Studio and be more than happy with it. So, tying it all back together, if you have to pick between the Pro and the Max chip, I think that it's a really uphill battle to justify getting the Max in the laptop anyway. in a desktop Mac Studio. Love it. Buy it all day long. In fact, I kind of wish that Apple had launched the Mac Studio as well because then it would have actually been like, "Hey, this is exciting. Go buy it." But now you're like, "Okay, it's $1,000 for this chip in a laptop that's about to get replaced." Like, is anybody even going to watch this video? I don't know. But either way, the point is, if you're buying a desktop, buy a Max chip in the Mac Studio. And if you're buying a laptop, get the Pro chip in a MacBook Pro. That's pretty much as simple as I can make it. So, I hope that you guys found this video helpful. Let me know what you think about the Max chips and about the new generation of MacBook Pro that's coming in just a couple of months. Of course, be sure to get subscribed if you don't want to miss any of that coverage. A big thanks to you guys for watching this video and I'll see you in the next one.

Thanks to Bellroy for sponsoring! Get 10% off at Bellroy using my link: https://bit.ly/4bhwNug The M5 Max is truly a crazy chip- this amount of performance in a laptop form factor is wild. But given that nobody is going to buy the MacBook Pro at this point with new ones coming in the fall, and the VERY high price tag, is M5 Max worth it? Time stamps 0:00 M5 Max is PRICEY 3:10 CPU performance 3:59 GPU performance 5:06 Gaming tests 7:36 AI 8:52 Video editing 10:17 Thermals and Battery 12:09 Is M5 Max worth it? Follow me on Tik Tok: https://www.tiktok.com/@lukemianiyt Follow me on Twitter: https://twitter.com/LukeMiani Gear: iFixit Pro Toolkit: https://amzn.to/3bYIzdM Sony A7Siii: https://amzn.to/3SWqtQG Sony ZV-E1: https://amzn.to/3VGoeCE Sennheiser MKH 416: https://amzn.to/31zGwt2 Zoom H6: https://amzn.to/2ECYSk6 Edelkrone SliderOne V2: https://amzn.to/33mTZmt *Amazon Links are affiliate codes and generate small commission to support the channel*