Comment by sho
8 hours ago
> On September 22, Apple is releasing the M5 Mac Studio with 256 GB of unified memory [..] it will probably [..] enough to write this snippet of code in 3 seconds
The author has obviously never ran an LLM on a mac! In 3 seconds, it will have possibly started to think about maybe scheduling a date to contemplate the planning timeline for processing the second token in your prompt.
The difference is memory bandwidth. The M5 Ultra that's coming out on 22nd September can do 1,200GB/s. The M5 Max you can buy today only has 614GB/s.
So, that gets us to about where nVidia was with Ampere in 2020. Let's hope the M7 catches us up with at least Hopper.
While true, the news here is the size of the unified RAM. Nvidia only exceeded 256GB RAM in the 2025 B300 - 288GB. The B300 alone (without the baseboard/PSU/chassis/wiring/CPUs/system RAM/etc) is at least 700% more expensive. This enables large language models on consumer hardware. 1200GB/s is plenty for many tasks.
2 replies →
It is and it isn't. Why are you comparing the m5max instead of the m4ultra?
The big deal to me is the number of compute cores for prefill tps, which is suppose to be 4x faster on the m5ultra.
It's my opinion that the m5 ultra is going to be a really big deal in terms of local AI accessibility. Flash sized models (~200-300b params) are going to be reasonably fast as long as you aren't throwing 40k context at it on each or the first request (ie, agentic harnesses).
Even agentic harnesses like Cline should move at a reasonable clip on m5 ultra. I suppose we will know sooner than later.
FYSA: Former m4 ultra 512GB owner and current 4x rtx6000 owner here. I upgraded because I needed more prompt processing speed and concurrency.
What do you use all that local tokens/second for?
The complaint is about prefill which is not memory bandwidth bound, it's compute bound. But they added neural accelerators for matmuls to the shader cores which should make prefill faster.
The author put in the numbers, but maybe you didn’t read them.
45 t/s a second is perfectly respectable especially with no limits and 24/7 uptime with very little power draw on the Studio.
Luna is at around 100 t/s for comparison, but it’s a worse model than 5.3 Flash
The joke is that macs are famously slow at prompt prefill and you are not getting anything back in 3 seconds, or probably even 30. Once they get generating, it can be acceptable, but the TTFT is horrendous.
There's a ton of well-understood things Apple can and hopefully will do to massively accelerate every stage of this pipeline and hopefully they're hard at work implementing most of them for m7.
> The joke is that macs are famously slow at prompt prefill and you are not getting anything back in 3 seconds.
Your knowledge is out of date. In truth it depends on the Mac and the models used.
I asked this question on M5 Max 128GB, using Ollama model Quen3.8:27b-mlx, with thinking enabled.
Question: "Give me a python code snippet that opens a file and sorts the lines of text. "
In 2.4 seconds it gave me 4 examples that work with different sorting configurations and a summary of when to use each.
Compare that to an older model of gpt-oss:20b, took 5 seconds to finish thinking and 2 seconds to stream the answer. It gave me one python example snippet and two one liners that do the same thing.
2 replies →
I think it's been pretty much proven by now that there are no cases where local inferencing is better than remote inferencing, unless absolute privacy is a hard requirement. The efficiencies that come with datacenter scale and hw can't be beaten.
Yeah, but data centers don't usually host abliterated models, hence the point of the article.
Also, LLMs never *write* code snippets, they just pirate them from somewhere else.
So, like humans? Code didn't just appear in my brain, I learnt it from reading it everywhere else.
Can you show me an example of a time that you prompted an LLM to provide some code, it did so, and then you were able to track down an original source for the output?
News: "LLM Models might kill us all!..."
HN user: "I take issue with the precise definition of one word in the article..."