> Testing conducted by Apple in January 2026 using preproduction 13-inch and 15-inch MacBook Air systems with Apple M5, 10-core CPU, 10-core GPU, 32GB of unified memory, and 4TB SSD, and production 13-inch and 15-inch MacBook Air systems with Apple M4, 10-core CPU, 10-core GPU, 32GB of unified memory, and 2TB SSD. Time to first token measured with an 8K-token prompt using a 14-billion parameter model with 4-bit quantization, and LM Studio 0.4.1 (Build 1). Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Air.
Oh dear 14B and 4-bit quant? There are going to be a lot of embarrassed programmers who need to explain to their engineering managers why their Macbook can't reasonably run LLMs like they said it could. (This already happened at my fortune 20 company lol)
And not even diehard Apple fanboys deny this.
I genuinely feel bad for people who fall for their marketing thinking they will run LLMs. Oh well, I got scammed on runescape as a child when someone said they could trim my armor... Everyone needs to learn.
There definitely are some who fit into this category, but if they're buying the latest and greatest on a whim then they've likely got money to burn and you probably don't need to feel bad for them.
Reminds me of the saying: "A fool and his money are soon parted".
That's how they make loot on their 128GB MacBook Pros. By kneecapping the cheap stuff. Don't think for a second that the specs weren't chosen so that professional developers would have to shell out the 8 grand for the legit machine. They're only gonna let us do the bare minimum on a MacBook Air.
Latency to the first token is not like a web page where first paint already has useful things to show. The first token is "The ", and you'll be very happy it's there in 50ms instead of 200ms... but then what you really want to know is how quickly you'll get the rest of the sentence (throughput)
Are they doubling down on local LLMs then?
I still think Apple has a huge opportunity in privacy first LLMs but so far I'm not seeing much execution. Wondering if that will change with the overhaul of Siri this spring.
I don't mind it, I open Apple stock. But I'm def not buying into their rebranding of integrated GPU under the guise of Unified Memory.
Aren't the OpenClaw enjoyers buying Mac Minis because it's the cheapest thing which runs macOS, the only platform which can programmatically interface with iMessage and other Apple ecosystem stuff? It has nothing to do with the hardware really.
Still, buying a brand new Mac Mini for that purpose seems kind of pointless when a used M1 model would achieve the same thing.
Yeah because Mac upgrade prices were already sky high, long before the component shortage. 32GB of DDR5-6000 for a PC rocketed from $100 to $500, while the cost of adding 16GB to a Mac was and still is $400.
But if the contract was for a specific amount of RAM and then people start coming to Apple more for high RAM machines, they're going to exhaust their contract sooner than usual and run out of cheap memory to buy. Then they have to decide if they want to lower their margins or raise the already-high price up to nosebleed levels.
[1]: https://developer.apple.com/documentation/virtualization/usi...
That's likely only part of the reason. Mac Mini is now "cheap" because everyone exploded in price. RAM and SSD etc have all gone up massively. Not the mention Mac mini is easy out of the box experience.
I considered the mac mini at the time, but the mac mini only makes sense if you need the local processing power or the apple ecosystem integration. It's certainly not cheaper if you just need a small box to make API calls and do minimal local processing.
If you just need "a small box to make API calls and do minimal local processing" you an also just buy a RPI for a fraction of the price of the GMKtec G10.
All 3 serve a different purpose; just because you can buy a slower machine for less doesn't mean the price:performance of the M1 Mac Mini changes.
Sadly not really. The Pi 5 8gb canakit starter set, which feels like a more true price since it's including power supply, MicroSD card, and case, is now $210. The pi5 8gb by itself is $135.
A 16gb pi5 kit, to match just the RAM capacity to say nothing of the difference in storage {size, speed, quality} and networking, is then also an eye watering $300
lol. you need to look at rpi 5 prices again. they are insane.
Do you really need Openclaw now? And not claude code + zapier or Claude code + cron?
That's the point. If you have worse CPU and GPU Windows will be sluggish (it's bloated).
This arb you’re talking about doesn’t exist. An m1 studio with 64 gb was $1300 prior to openclaw. You’re not getting that today.
I would have preferred that too since I could Asahi it later. It’s just not cheap any more. The m4 is flat $500 at microcenter.
For the same price in API calls, you could fund AI driven development across a small team for quite a long while.
Whether that remains the case once those models are no longer subsidized, TBD. But as of today the comparison isn't even close.
Assuming, of course, that your legal team signs off on their assurance not to train on or store your data with said Enterprise plans.
With Anthropic you're paying for "more tokens than the free plan" which has no meaning
To be clear, I totally get the idea of running local LLMs for toy reasons. But in a business context the sell on a stack of Mac Pros seems misguided at best.
It is the first local model I've tried which could reason properly. Similar to Gemini 2.5 or sonnet 3.5. I gave it some tools to call , asked claude to order it around, (download quotes, print charts, set up a gnome extension) even claude was sort of impressed that it could get the job done.
Point is, it is really close. It isn't opus 4.5 yet, but very promising given the size. Local is definitely getting there and even without GPUs.
But you're right, I see no reason to spend right now.
I've been working my way up from a 3090 system and I've been surprised by how underwhelming even the finetunes are for complex coding tasks, once you've worked with Opus. Does it get better? As in, noticeably and not just "hallucinates a few minutes later than usual"?
> But I'm def not buying into their rebranding of integrated GPU under the guise of Unified Memory.
But it is Unified Memory? Thanks to Intel iGPU term is tainted for a long time.
And while it is stupid slow, you can run models of hard drive or swap space. You wouldn’t do it normally, but it can be done to check an answer in one model versus another.
Sonnet is so fast too. GPT-5.2 needs reasoning tuned up to get tool calling reliable and Qwen3 Coder Next wasn’t close. I haven’t tried Qwen3.5-A3B. Hearing rave reviews though.
If you’re using successfully some model knowing that alone is very helpful to me.
Now extrapolating in line with how Sun servers around year 2000 cost a fortune and can be emulated by a 5$ VPS today, Apple is seeing that they can maybe grab the local LLM workloads if they act now with their integrated chip development.
But to grab that, they need developers to rely less on CUDA via Python or have other proper hardware support for those environments, and that won't happen without the hardware being there first and the machines being able to be built with enough memory (refreshing to see Apple support 128gb even if it'll probably bleed you dry).
Basically, too many choices to "focus on" makes non a winner except the incumbent.
I certainly only use Macs when being project assigned, then there are plenty of developers out there whose job has nothing to do with what Apple offers.
Also while Metal is a very cool API, I rather play with Vulkan, CUDA and DirectX, as do the large majority of game developers.
Only groups of developers more tied to Windows that I can think of are probably embedded people tied due to weird hardware SDK's and Windows Active Directory dependent enterprise people.
Outside of that almost everyone hip seems to want a Mac.
Everyone hip alright, or at least those that would dream to earn a salary big enough to afford Apple taxes.
Remember there are world regions where developers barely make 1 000 euros per month.
https://survey.stackoverflow.co/2025/technology/#1-computer-...
The US 1s? Is that why we have Deepseek and then other non-US open source LLMs catching up rapidly?
World view please. The developer community is not US only.
It wouldn’t surprise me if the deepseek people were primarily using Mac’s. Maybe Alibaba might be using PCs? I’m not sure.
People always overlook that CUDA is a polyglot ecosystem, the IDE and graphical debugging experience where one can even single step on GPU code, the libraries ecosystem.
And as of last year, NVidia has started to take Python seriously and now with cuTile based JIT, it is possible to write CUDA kernels in pure Python, not having Python generate C++ code that other tools than ingest.
They are getting ahead of Modular, with Python.
Neural Accelerators (aka NAX) accelerates matmults with tile sizes >= 32. From a very high level perspective, LLM inference has two phases: (chunked) prefill and decode. The former is matmults (GEMM) and the latter is matrix vector mults (GEMV). Neural Accelerators make the former (prefill) faster and have no impact on the latter.
I assume they have a moderate bet on on-device SLMs in addition to other ML models, but not much planned for LLMs, which at that scale, might be good as generalists but very poor at guaranteeing success for each specific minute tasks you want done.
In short: 8gb to store tens of very small and fast purpose-specific models is much better than a single 8gb LLM trying to do everything.
Apple is in the hardware business.
They want you to buy their hardware.
People using Cloud for compute is essentially competitive to their core business.
Remains to be seen how capable it actually is. But they're certainly trying to sell the privacy aspect.
It's the best. We all turned it off. 100% privacy.
Are they doubling down on local LLMs then?
Apple's stated AI strategy is local where it can and cloud where it needs. So "doubling down"? Probably not. But it fits in their strategy.
I think I'll pass on upgrading.
Honestly, I think that's the move for apple. They do not seem to have any interest in creating a frontier lab/model -- why would they give the capex and how far behind they are.
But open source models (Kimi, Deepseek, Qwen) are getting better and better, and apple makes excellent hardware for local LLMs. How appealing would it be to have your own LLM that knows all your secrets and doesnt serve you ads/slop, versus OpenAI and SCam Altman having all your secrets? I would seriously consider it even if the performance was not quite there. And no need for subscription + cli tool.
I think apple is in the best position to have native AI, versus the competition which end up being edge nodes for the big 4 frontier labs.
I love the push to local llms. But it’s hilarious how apple a few years ago was so reluctant to even mention “AI” in its keynotes and fast forward a couple years they’ve fully embraced it. I mean I like that they embraced it rather than be “different” (stubborn) and stay behind the tech industry. It’s the smart choice. I just think it’s funny.
"AI" (LLMs) may or may not have a bubble-pop moment, but until it does Apple get to ride it on these press releases and claims. But if the big-pop occurs, then Apple winds up with really fantastic hardware that just happens to be good at AI workloads (as well as general computing).
For example, image classification (e.g. face recognition/photo tagging), ASR+vocoders, image enhancement, OCR, et al, were popular before the current boom, and will likely remain popular after. Even if LLM usage dries up/falls out of vogue, this hardware still offers a significant user benefit.
What is more likely to happen though is that it doesn't take multiple $10B of datacenter and capital to build out models--and the performance against LLM benchmarks starts to max out to the point where throwing more capital at it doesn't make enough of a difference to matter.
Once the costs shrink below $1B then Apple could start building their own models with the $139B in cash and marketable securities that they have--while everyone else has burned through $100B trying to be first.
Of course the problem with this strategy right now is that Siri really, really sucks. They do need to come up with some product improvements now so that they don't get completely lapped.
Unified memory is a serious architectural improvement.
How many GPUs does it take to match the RAM, and make up for the additional communication overhead, of a RAM-maxed Mac? Whatever the answer, it won’t fit in a MacBook Pro’s physical and energy envelopes. Or that of an all-in-one like the Studio.
So yes, the LLM should be inferencing on your prompt, but it should also be inferencing on 25,000 other things … in parallel.
Those are the compute needs.
We just need compute everywhere as fast as possible.
I just don't get why they're dropping the ball so much on this.
They aren’t dropping the ball, they are being smart and prudent.
Do think it'll be common to see pros purchasing expensive PCs approaching £25k or more if they could run SoTA multi-modal LLMs faster & locally.
So as most people in or adjacent to the AI space know, NVidia gatekeeps their best GPUs with the most memory by making them eye-wateringly expensive. It's a form of market segmentation. So consumer GPUs top out at 16GB (5090 currently) while the best AI GPUs (H200?) is 141GB (I just had to search)? I think the previou sgen was 80GB.
But these GPUs are north of $30k.
Now the Mac Studio tops out currently at 512GB os SHARED memory. That means you can potentially run a much larger model locally without distributing it across machines. Currently that retails at $9500 but that's relatively cheap, in comparison.
But, as it stands now, the best Apple chips have significantly lower memory bandwidth than NVidia GPUs and that really impacts tokens/second.
So I've been waiting to see if Apple will realize this and address it in the next generation of Mac Studios (and, to a lesser extend, Macbook Pros). The H200 seems to be 4.8TB/s. IIRC the 5090 is ~1.8TB/s. The best Apple is (IIRC) 819GB/s on the M3 Ultra.
Apple could really make a dent in NVidia's monopoly here if they address some of these technical limitations.
So I just checked the memory bandwidth of these new chips and it seems like the M5 is 153GB/s, M5 Pro is ~300 and M5 Max is ~600. I was hoping for higher. This isn't a big jump from the M4 generation. I suspect the new Studios will probably barely break 1TB/s. I had been hoping for higher.
5090 has 32GB, and the 4090 and 3090 both have 24GB.
I also haven’t seen any improvements in the frontier models in years, and I’m anxiously awaiting local models to catch up.
This correlation of Apple and privacy needs to rest. They have consistently proven to be otherwise - despite heavily marketing themselves as "privacy-first"
https://www.theguardian.com/technology/2019/jul/26/apple-con...
Do I wish Apple was way more transparent and gave users more control over gatekeeper and other controversial features that erode privacy? Absolutely.
> Even More Value for Upgraders
> The new 14- and 16-inch MacBook Pro with M5 Pro and M5 Max mark a major leap for pro users. There’s never been a better time for customers to upgrade from a previous generation of MacBook Pro with Apple silicon or an Intel-based Mac.
I read as "Whoops we made the M1 Macbook Pro too good, please upgrade!"
I think I will get another 2-5 years out my mine.
Apple: If you document the hardware enough for the Asahi team to deliver a polished Linux experiene, I'll buy one this year!
for example, let's say the new os depends on m5's exclusive thumbnail generator accelerator, and let's say it improves speed by a 20%.
now, your M1 notebook than on previous OSes uses standard gpu acceleration for thumbnails will not have this specialized hardware acceleration, it will have software fallback that will be 90% slower.
you won't notice it a first thought because it's stuff, fast, but it eats a bit of the processor.
multiply this by 1000 features and you have a slow machine.
I don't know how else to explain how an ipad pro cannot even scroll a menu without stuttering, it's insane how fast these things were on release
the Liquid Glass for example probably is not so great when it comes to resources. Probably works better with latest metal and hardware blocks on the GPU in M5 as opposed to using GPU cores and unified memory on 8gb M1 making latest macOS work not so great. I have the M1 8gb air and it is really slow on Tahoe. It was snappy just a couple of years ago on a fresh install.
Not upgrading any of my Macs ever again. I was a fanboy looking for every new update like a present, for 13 years, not anymore. It took one Tahoe burn all that trust. Never upgrading major OS versions on hardware from Apple again.
I've been holding out as you do for as long as I can but in 1-2 years the apps just stop working (some of them).
Win 11 is bad compared to Win 10 as well. I'm fairly new to Linux so I can't really form an opinion there.
My next MBP will have 128GB memory, but these prices just wanna make me wait longer.
Nothing has broken and I consistently get 4-6 hours of heavy work time while on battery. An amazing machine for the price I paid.
As there target for that marketing, I can report it hits home!
But objectively, there is nothing wrong with my current experience at all.
I have never had that experience over many generations and types of machines. The M1 keeps looking better and better in hindsight.
—-
Looking forward, either the M5 is the next M1, a bump of good that will last. Or Apple will be really firing on all cylinders if it can “obsolete” the M5 anytime soon.
~9 years later, there are a lot of people still using it as their main machine, waiting until we get kicked off the corp network for lack of software support.
It feels really stupid to have to throw away a perfectly capable machine with 64GB of RAM in 2026.
Barring removal of Esc key, I think the touch bar was useful because it showed contextual actions. But not every app used it so it didn't really get a chance to shine.
I think people that do do tasks where a touch screen makes sense are probably just doing most of their work on an iphone or an ipad anyway.
Now gesture control on VR/AR setups? Sure, that feels like a new human/computer interaction system that makes sense. Jabbing at my laptop screen with one hand on my keyboard, not so much.
I have a M1 MacBook Pro with the touch bar since. It’s crap. I remember the keynote where they introduced it and a DJ mixed music using it. It was ridiculous that it got approved.
Fortunately I just keep my laptop closed and use an attached display and keyboard and mouse, so I don't even remember if my M1 has a touch bar.
Also minor nit: it's seldom, not seldomly. Seldom certainly doesn't seem like an adverb, but it is.
I still don't have a strong urge to upgrade. I could probably get by on 32GB (like my work-issued machine is) but 64GB is the right amount of headroom for me.
>I think I will get another 2-5 years out my mine.
I only own a M4 because the M1 had a hardware fault and I needed a replacement ASAP. (I sold the M1 after repair.)
Although I'm glad to have a newer machine with longer future support, I have yet to notice any meaningful performance difference.
the air series is really good, and very light
my M1 is now noticeably heavy and I don't think upgrading to another Macbook Pro is the move the resell value of the M1 did not hold, specifically the bumped up storage models. There doesn't seem to be a market for 8TB of space specifically, but the base 1 - 2TB holds its value because the baseline of the MBP holds its value
M5 Max looks tempting if there is a very compelling tradein, but the M1 Max is pretty old so I don't have real hope of that, but I'll look. For AI Inference the difference doesn't seem good enough yet and necessary enough. I'll still need to use the cloud or aspire to have a specialized machine with more RAM or circuitry on my network.
Interested to see what FP32 values they have for a site I've been working on [0].
[0]: https://flopper.io
Interestingly, 36-128GB models are showing as “currently unavailable” on the store page, and you can’t even place an order for them right now? But for anyone curious, it’s quoting $5099 for the 128GB RAM 14” MacBook Pro model.
No change from the previous models then, 16GB->32GB was already $400. They're cutting into their previously enormous margins to keep the prices stable, rather than hiking the prices to maintain their margins.
I wonder if there's a fab time secondary market where Wall Street types are making millions off speculating fab time.
Most people can totally live with 16gigs but it is kind of a waste for the horsepower. They know what they are doing. Apple is a master in upselling.
Though personally I don't mid the aggressive upsellign as long as the quality is there. Problem is, the hardware quality is great but the software side is severely lacking and getting worse.
Which, I mean, I love unified memory, as one of those weirdos that does do local LLM stuff and am contemplating if it's time to upgrade my m2 max.
But if you needed 32gb then you still need at least 32gb now. Unless swap on nvme disks is enough for you - and it isn't for me.
Isn't this it?
My M3 Pro from a few years ago for the same price had 18GB.
Interesting that this hasn't budged since the memory shortages appeared.
Apple has had enough war chests with the ability of buying the entirety of TSMC's new capacity years in advance in the past.
If I were to guess, Apple locked in their entire BOM and production capacity two years ago. That's something even the large players cannot replicate because they run cash-lean and have too many different SKUs, and the small players (Framework, System76, even Steam) are entirely left to the forces of the markets.
I'd love to have customers like Apple. Bumps $200: "it didn't change!!!"
And no power adapter included.
Now it starts at $1699, a $100 bump but comes with a 1TB SSD. Previously it would have cost $1799 for the 1T SSD, so it's a $100 bump on base price but you are also getting 1TB SSD for $100 less than before.
Like for my last upgrade, I bit the bullet and upgraded to 1TB for the first time ever instead of base storage at Apple's absurd prices, so it's good, but if I'd not have been willing to spend money on that at all, they lifted the floor.
My cell phone plan has been increasing every year by small amounts, but my usage pattern hasn't changed, and meanwhile they've restricted HD streaming using Deep Packet Inspection or whatever, so I theoretically have a 100GB full speed cap but can't practically use more than 20gb anyway, so they're pricing the bandwidth into the contract but I can't save money by getting a lower ceiling
Try making a good product that people love?
To be fair, ever since the advent of high power USB-C PD that really, really is not needed any more, way too many power bricks are effectively e-waste.
People already have USB-C power bricks and docks everywhere and unlike pre-USB-C generations, you can use them not just across different generations of hardware, but across vendors as well.
Which MacBooks have been for almost a decade - the 2016 MBP with Touch Bar was the first that went fully USB-C PD. Anyone who has had a MacBook in that time frame will have had at least one high power USB-C PD wall wart.
The Windows world, as usual, has been different, but even there, I'm not aware of any mainstream model being sold in the last two years without even a single PD capable port.
I wonder if that would happen regardless of RAM, e.g. for tariffs etc.
Here in the UK, they no longer include the power adapter even with the top models. I just specced out a fully-loaded M5 Max Macbook Pro, 128GB RAM, 8TB storage on the Apple Store, and it doesn't include a power adapter by default.
The 140W power adapter can be added as an option to the MacBook Pro for an additional £99 + VAT, or purchased separately. If you purchase separately you can of course choose a lower-power adapter for a lower price.
Now that a power adapter isn't included and you have to pay for it separately, it might make more sense to get one of the good brands of GaN power adapters instead, because they are smaller than the Apple ones for the same power, and have more ports.
That's incredibly stupid(of apple), I'm in the UK and literally got my M4 Max MacBook Pro delivered on Friday, it came with a power adapter.
this is my exact opposite experience. my M3 Max from 2 years ago now has <2hrs battery life at best. wondering if any experts here can help me figure out what is going on? what should i be expecting?
I type this from an M3 Max 2023 MBP that still has 98% battery health. But admittedly it's only gone through 102 charge cycles in ~2 years.
(use `pmset -g rawbatt` to get cycle count or `system_profiler SPPowerDataType | grep -A3 'Health'` to get health and cycles)
Apple will replace the battery for $249 if you choose to. https://support.apple.com/mac-laptops/repair?services=servic...
I'm typing this on an M3 Max; its max battery capacity is 88%. I've got some things running (laptop average temp is 50-55C, fans off), screen is half brightness, and it's projected to go from 90% to 0% in five hours. I don't usually baby it enough to test this, but 8-10 hours should be achievable.
And some apps are really inefficient. New Codex app drains my battery. If you are using Codex I recommend minimizing it, since it’s the UI that uses most power.
What in the world is an idle Claude Desktop doing that uses so much power?
Hot take: people should get used to, and expect to, replace device batteries 1 or 2 times during the device lifetime. They're the main limiting factor on portable device longevity, and engineers make all kinds of design tradeoffs just to make that 1 battery that the device ships with last long enough to not annoy users. If we could get people used to taking their device in for a battery once every couple of years, we could dramatically reduce device waste, and also unlock functionality that's hidden behind battery-preserving mechanisms.
I’m not even sure how it got installed, possibly when I installed Zoom for an interview once but I don’t know. Point is, at least in one case, AI can help track down battery hogs.
Incidentally, I just switched to Asahi Linux, but that was for software quality and openness reasons, rather than anything to do with performance.
Unfortunately, number always must go up (and the rate at which the number goes up, also must go up).
1. While the hardware and performance are amazing, the user interface is the opposite. Imagine buying a luxury car with amazing performance only to find that simply opening the door is a royal pain, each and every time.
2. Apple will downgrade the usability over time. A year from now, or two, Apple will downgrade your user experience. Imagine that in your luxury car you can see out the windshield, but the dealer insists that you install a new upgrade with a heads-up-display that cannot be turned off.
3. Apple will degrade the performance of your system over time by constantly introducing more features which require better hardware. Your sleek and fast computer will eventually become unusably slow.
4. Apple profits from preventing you from using the computer you own with other software, for example Linux. When your computer cannot run Mac OS (see #3) above or you get sick of the "features" (see #1 and #2 above), you will not be able to do so. The reason for this is if you could try Linux, there is is a strong possibility you will see just how user unfriendly Mac OS is and never go back.
5. You care about the environmental impact of your purchasing decisions. You understand that because you are not able to upgrade the hardware and operating system, your purchase is very likely to end up in a landfill.
I use my laptop for development. I don't actually use most of the built in applications. My browser is Firefox, I use codex, vs code, intellij, iterm2, etc. Most of that works just fine just as it did on previous versions of the OS. I actually on purpose keep my tool chains portable as I like to have the option to switch back to Linux when I want to. I've done that a few times. I come back for the hardware, not the OS.
In my experience, if you don't like Apple's OS changes that is unfortunate but they don't seem to generally respond to a lot of the criticism. Your choices are to get further and further out of date, switch to something else, or just swallow your pride. Been there done that. Windows is a "Hell No" for me at this point. I'll take the UX, with all the pastel colors that came and went and all the other crap that got unleashed on macs over the last ten years. Definitely a case of the grass not being greener on Windows. Even with the tele tubby default desktop in XP back in the day.
I can deal with Linux (and use that on and off on one of my laptops). However, that just doesn't run that well on mac hardware. And any other hardware seems like a big downgrade to me. Both Windows and Linux are arguably a lot worse in terms of UX (or lack thereof). Linux you can tweak. And you kind of have to. But it just never adds up to consistent and delightful. Windows, well, at this point liking that is probably a form of Stockholm Syndrome. If that doesn't bother you, good for you.
So, Mac OS it is for me as everything else is worse. I've in the past deferred updates to new versions of Mac OS as well. Generally you can do that for a while but eventually it becomes annoying when things like homebrew and other development toys start assuming you run something more recent. And of course for security reasons you might just not drag your feet too long. Just my personal, pragmatic take.
Literally unusable
I already left the beta train on my iPhone because I had too many issues getting my grocery apps to allow me to place orders without going to my laptop and doing it in a web browser.
It's so bad I switched back to Chrome. I had thought Chrome had a major battery life penalty compared to Safari on Macs, but I checked more up-to-date info and apparently that's outdated.
But I think this predates Tahoe.
The new tensor cores, sorry, "Neural Accelerator" only really help with prompt preprocessing aka prefill, and not with token generation. Token generation is memory bound.
Hopefully the Ultra version (if it exists) has a bigger jump in memory bandwidth and maximum RAM.
Most stuff ends up running Metal -> GPU I thought
https://creativestrategies.com/research/m5-apple-silicon-its...
That's actually the biggest growth area in LLMs, it is no longer about smart, it is about context windows (usable ones, note spec-sheet hypotheticals). Smart enough is mostly solved, combating larger problems is slowly improving with every major release (but there is no ceiling).
This seems even likely as the memory bandwidth hasn't increased enough for those kinds of speedups, and I guess prefill is more likely to be compute-bound (vs mem bw bound).
I'd take that tradeoff. On my M3 Ultra, the inference is surprisingly fast, but the prompt processing speed makes it painful except as a fallback or experimentation, especially with agentic coding tools.
Wondering if local LLM (for coding) is a realistic option, otherwise I wouldn't have to max out the RAM.
For reference:
| model | size | params | backend | threads | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | ------: | --------------: | -------------------: |
| qwen35 ?B Q5_K - Medium | 6.12 GiB | 8.95 B | MTL,BLAS | 6 | pp512 | 288.90 ± 0.67 |
| qwen35 ?B Q5_K - Medium | 6.12 GiB | 8.95 B | MTL,BLAS | 6 | tg128 | 16.58 ± 0.05 |
| model | size | params | backend | threads | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | ------: | --------------: | -------------------: |
| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | MTL,BLAS | 6 | pp512 | 615.94 ± 2.23 |
| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | MTL,BLAS | 6 | tg128 | 42.85 ± 0.61 |
Klein 4B completes a 1024px generation in 72seconds.I think the truth is somewhere in the middle, many people don't realize just how performant (especially with MLX) some of these models have become on Mac hardware, and just how powerful the shared memory architecture they've built is, but also there is a lot of hype and misinformation on performance when compared to dedicated GPU's. It's a tradeoff between available memory and performance, but often it makes sense.
Also, the mix of cores have changed drastically.
- 6 "Super cores"
- 12 "Performance cores"
I'm guessing these are just renamed performance and efficiency cores from previous generations.
This is a massive change from the M4 Max:
- 12 performance cores
- 4 efficiency cores
This seems like a downgrade (in core config but may not be in actual MT) assuming super = performance and performance = efficiency cores.
I think this is a new design, with Apple having three tiers of cores now, similar to what Qualcomm has been doing for a while.
I think how it breaks down is:
- "Super" are the old "P" cores, and the top tier cores now
- "Performance" cores are a new tier and seen for the first time here, slotting between "old" P and E in performance
- "Efficiency" / "E" are still going to be around; but maybe not in desktop/Pro/Max anymore.
For example, 6 super, 8 performance, and 4 efficiency.
Whoah, both the Pro and Max CPUs feature 18 cores. This hasn't happened since M1 Pro/Max. This is a surprise.
The base M5 has super/efficiency cores.
The Pro and Max have super/performance cores.
I believe they lower the clock speed, limit how much work is done in parallel on each core, and limit how aggressive the speculative execution is so less work is wasted.
The M5 performance cores can be scaled down to match efficiency cores in performance and power usage.
> The industry-leading super core was first introduced as performance cores in M5, which also adopts the super core name for all M5-based products
But new "performance" is claimed to be new design (= not just overclocked efficiency core from M5?):
> M5 Pro and M5 Max also introduce an all-new performance core that is optimized to deliver greater power-efficient, multithreaded performance for pro workloads.
quotes from https://www.apple.com/newsroom/2026/03/apple-debuts-m5-pro-a...
> The tech giant says the chips are engineered around its new Fusion Architecture, an advanced design that merges two dies into a single, high-performance system on a chip (SoC), which includes a powerful CPU, scalable GPU, Media Engine, unified memory controller, Neural Engine, and Thunderbolt 5 capabilities.
https://techcrunch.com/2026/03/03/apple-unveils-m5-pro-and-m...
They also replaced the efficiency cores on the CPU chiplet with a new higher performance design.
> The CPU now features six “super cores,” which is Apple’s term for its highest-performance cores, alongside 12 all-new performance cores. Collectively, the CPU boosts performance by up to 30% for pro workloads.
I have not once felt the need to upgrade in years, and that’s with doing pretty demanding 3D and LLM work.
The high memory Macs have been great for being able to run LLMs, but the prompt processing has always been on the slow side. The new AI acceleration in these should help with that.
There are also workloads like compiling code where I’ll take all the extra speed I can get. Every little bit of reduced cycle time helps me finish earlier in the day.
And then there’s gaming. I don’t game much, but the M1 and M2 era Apple Silicon feels sluggish relative to what I have on the nVidia side.
and that’s with doing pretty demanding 3D and LLM work.
M5 Max maxes out at 128GB, so that will have to wait for the eventual M5 Ultra anyways.
Even if a new device is a small upgrade from last year's model, it can be a giant upgrade for other people.
It's the first time I've ever been so repulsed by a design that I actively avoid it just... out of sheer preference.
If you move your home directory to a different disk partition, you can even share it between two different macOS versions!
I have a Intel-based 2019 Macbook Pro still and I have NEVER in its lifetime gotten even half of what they are claiming here. These days if I run it from battery I might get 90 mins.
That said I had a maxed out Macbook Pro M4 Max on order but just cancelled it right now and will get this new M5 Max one for basically the same price. Once I saw that they didn't up the price of memory (I don't know how it doesn't affect them) I canceled my order.
CoreImage - GPU accelerated image processing out of the box;
ML/GPU frameworks - you can get built-in, on device's GPU running ML algorithms or do computations on GPU;
Accelerate - CPU vector computations;
Doing such things probably will force you to have platform specific implementations anyway. Though as you said - makes sense only in some niches.
I won’t try to claim that Electron and friends have no place is software development but we absolutely should be pushing back harder against stuffing it everywhere it possibly can be.
Every modern desktop uses webviews in some capacity. macOS renders many apps with webviews, GNOME uses gjs to script half the desktop. The time to push back was 10-20 years ago, it's too late to revert now.
I don't think gjs is a webview. It uses JavaScript, granted, but binds to a native toolkit, not to DOM and CSS.
Meanwhile on Windows major features like the Start menu are written in React.
Worth noting that WebKit webviews also tend to be more lightweight than their Chromium brethren.
I think I read somewhere long time ago that Capture One is also using Qt for GUI, though cannot find this anymore, so probably not true.
Now it starts at $1699, a $100 bump but comes with a 1TB SSD. Previously it would have cost $1799 for the 1T SSD, so it's a $100 bump on base price but you are also getting 1TB SDD for $100 less than before.
For example, up until MacBookPro M2, MacBookPro M2 came with M2 Pro chip.
However, starting with M3, Apple lowered the MacBookPro MSRP to $1599, but its base configuration was downgraded to M3 chip from M3 Pro. To get the M3 Pro, you had to pay $1999. There's substantial performance between the two.
Same with M4. To get the M4 Pro chip, you had to pay $1999.
Now to get M5 Pro chip, it's $2199. Still a good value, but just saying it's a deviation from the trend.
The prompt processing sped up.
Not the output generation.
M4 was notoriously slow at this compared to DGX etc.
For example, grab yourself an Omen Transcend 14, spec it to 64GB RAM and the RTX 5070. You’re under $2000 and getting better graphics performance for anything that isn’t AI, and you’ve got an upgradable 1TB SSD and removable WiFi card.
You’re also getting an OLED screen which most people would prefer.
This model in particular I’ve chosen because it’s just as quiet as the M4 MacBook Pro models within 3dB during high intensity usage and gets very similar battery life, actually better battery life than the M4 Pro/Max models for light tasks.
For those of us with astigmatism it's really night and day experience.
MacBook Pro with M5 Pro now comes standard with 1TB of storage, while MacBook Pro with M5 Max now comes standard with 2TB. And the 14-inch MacBook Pro with M5 now comes standard with 1TB of storage.This is the important statement. 614GB/s is quite decent, however a NVIDIA RTX 5090 already offers 1,792 GB/s (roughly 3x) of memory bandwidth, for comparison.
You can buy two m5 pro base model for the same price as a single 5090...
In Europe I can get a 128gb mac studio m4 max for 300 euros more than a 5090 (for which you still need to buy a power supply, motherboard, cpu , &c.)
TL;DW: 2010s intel mac era laptops have seen at very best 35% single core CPU performance over in 5 years time! This happens almost every year now with M line macs.
Rant:
Retina macs were great and had great form factor over unibody macs. Touch-bar macs in the mid 2010s was IMHO a disaster. Terrible keyboard, poorer thermal capacity, missing essential ports, adapters galore.
But when it comes to performance - early 2010s macbooks with dedicated gpus had serious overheating issues.
Retina macbooks were decent, both form factor and performance.
Touch-bar macs were totally abysmal, all performance gains over previous generations was all through pumping more heat. CPUs constantly pegged at 90C+, cannot have laptop on your lap, Apple planning and delaying release schedules around intel fumbling their tik/tok cycles (as far as i remember some macs did not get any improvements for 2 years+ if not way more). Upgrades sometimes were total jokes, because of thermal throttling there was no point to put more hardware than it could work with. From reviews buying higher level cpu sometimes didn’t give noticeable real life gains because, again, thermal throttling kicking in instantly. 2020 intel macbook pro has fans spinning almost all the time. Having a remote call - your battery is dead in 2h max (essentially 1% per 1min).
M1 mac gave insane perceived performance boost - no noticeable throttling. Macbook airs are fully passively cooled, never heard M Macbook pro with fans screeching.
Also real full work day battery doing real work without power adapter at full performance. Cool to touch most of the time.
I made homework for a job in 2020 on a 2013 personal macbook. Apart from memory footprint - I could not feel noticeable difference on development experience. Editing images was frustrating on both. With M macs - its silent, smooth fast.
Number of parallel cores matching best intel cpus on base models, GPU blowing any mobile gpu in price range out of the water with thermal capacity to peg it 100% no problem. Unified memory for those GPUs to do what you could only imagined doing on GPUs that cost 3 times more than the macbook.
It’s a such excellent architecture that yeah - it’s “boring” you can nitpick about M69 Ultra Pro Max performance, but take a base MBP of any M line and it blows almost any laptop out of the water, even to this day.
Part of this has to be blamed on Apple, as the chassis designer and system integrator. Intel did not force them to put an i9 in the 16" MBP, Apple made that decision. Even now, Apple refuses to use the old Touchbar chassis for anything other than passively-cooled base model chips. It's a tacit admission that they know the design failed; it probably would still suck with Pro and Max chips inside them.
The paper-thin unibody, Butterfly keyboard and Touchbar were all unpopular features, but Apple shipped them anyways. It really shouldn't take 4+ years to respond to critical design flaws, especially if you're a trillion-dollar business.
Phones have less configurability, they sell more, and colors seem more important.
It doesn't even look like they added cellular as an option with their own C1X chip (getting around the licensing / cost issues since it's their own chip now).
How is that different from the silicon interposer they were using before?
The big change is the two dies don’t have to fabbed next to each other in a single wafer, which is fantastic for costs and yields. But would this affect the interconnect speed somehow?
How would the two be wired together?
Could this mean the Ultra comes back in M6 since it would be easier to fab?
It's chiplets just like GB10, Strix Halo, etc. One die has the CPU and the other die has the GPU.
How is that different from the silicon [bridge] they were using before?
It's probably similar.
the two dies don’t have to fabbed next to each other
They never were; this is a widespread misunderstanding.
But would this affect the interconnect speed somehow?
Apple never documented the internal interconnect for the M4 Pro/Max and now they don't document it for the M5 Pro/Max so we don't know. It's probably better to read reviews and avoid theorycrafting and backseat driving.
They seem to market it as a technological advancement, which it is, but rather than being excited im actually worried about hidden latencies that could come with that approach. Have you found any interesting info on that yet?
Yes, I'm sure by then there will be better models on offer via cloud providers, but idk if I'll even care. I'm not doing science / research or complex mathematical proofs, I just want a model good enough to vibe code personal projects for fun. So I think at that point I'll stop being a OpenAI / Anthropic customer.
I think at this point Apple will just release new versions of laptops whenever new CPU revisions and yields allow. M5 Pro wasn't ready for October so delayed until now.
And another rumor said these are going to be updated again this fall but I’m not sure about that. With OLED screens and M6 (supposedly).
They also probably had RAM contracts in place far enough in advance to avoid the worst of the price spikes.
Wish it was Blender though ;)
Here in Europe, including 21% VAT, that's €6.124,00 ($7.094,35 equivalent).
Because of pricing strategies and such, the 128GiB version comes with a 2TiB SSD at minimum, and also requires the M5 Max (not Pro) at its highest configuration.
Not sure if this is new, but it should be noted that these laptops don't come with a charger any more.
70W USB-C Power Adapter (included with M5 Pro with 16-core GPU)
96W USB-C Power Adapter (included with M5 Pro with 20-core GPU, configurable with M5 Pro with 16-core GPU)
USB-C to MagSafe 3 Cable (2 m)Devices should be offered without a charger. There's no law that states that that should be the default configuration. Nor that the charger should cost extra.
It is interesting to see how mass-propaganda is playing out right before our eyes...
Linux in a VM would work with the usual caveats. Periphery like the built-in webcam most likely won't work. Getting codecs and DRM to run will be pain and you'll be back to use macOS for that quickly (but that's just standard pain of ARM Linux).
Touch ID is the least of the problems, but the other two are more serious.
It's one of those things, yes if I'm spending that much on a laptop I can afford to spend $80 on the adapter too, but does it feel good as a customer to do that or are you souring the experience of buying from you just to earn a few more dollars.
https://appleinsider.com/articles/25/10/15/eu-gets-what-it-a...
In the US they provide one in the box free of charge.
https://9to5mac.com/2025/10/16/no-the-eu-didnt-ban-apple-fro...
The EU requires that users must be able to buy a device without a charger. It's a huge supply chain challenge to add two variants of every single SKU, one with a charger and one without. So the obvious solution is to sell the charger separately, since you need that regardless, and always sell the device without a charger. You avoid having two variants of everything that way.
Now, you could maybe argue that Apple should default to bundle a charger with your laptop, so that you'd have to uncheck a "bundle charger" checkbox on their website. But do you really care whether your laptop costs $2200 and you can buy a charger for $60 or your laptop costs $2260 and you can save $60 by removing the charger?
You can make an argument that doing it Apple's way hides a price increase. And yeah, that's probably fair. But it's not like Apple is afraid of non-hidden price increases either.
So, if you want one of mine, you can have one. On me. Because I'm fucking drowning in the things and appreciate not having to deal with another one.
Which roughly translates to 30B Q8 size LLM at 10t/s for the M5 Pro and 60B Q8 size LLM at 10t/s for the M5 Max
For reference, RTX 3090 24GB has a memory bandwidth of approx. 936.2 GB/s, DGX Spark 128GB features a unified memory bandwidth of up to 273 GB/s
Might need to wait for the M5 Ultra or M6 Max with 128GB of RAM until the memory bandwidth is greater than a GTX 5090.
for reference, the M1 Max has 400GB/s of memory bandwidth, half a decade ago
I'm really wanting to build proper local-first AI workflows at home, and I think Apple has an opportunity to make that possible in a way other companies aren't really focused on, but we need significantly larger memory capabilities to do it, which I know is tough in the current memory market but should be available for a cost.
128 GB maximum.
Sigh.
You can run open source models like Kimi K or Qwen locally. Apple recently updated Xcode 26.3 to support local models.
My current work laptop (Lenovo) is quite a beast as well when plugged in, but I can literally see the battery percentage tick down while unplugged, but colleagues with their Macs can go all day.
Actually, I can think of one hardware want: have they gotten it to where you can do external GPUs and the like more easily?
Would still buy one over any other laptop on the market today for what I use them for.
LOL. is it repairable? probably not.
Just about to be time for me to get a new laptop. Typically I buy a generation behind, but want to make sure I won't miss anything huge.
https://www.bloomberg.com/news/articles/2026-02-24/apple-s-t... (https://archive.ph/qT3QV)
This is just marketing speak. Stop repeating marketing. It isnt a walled garden, its a walled prison.
Unified memory is just regular memory. There is nothing special about integrated GPUs.