Spent a week trying to get sensible results out of llama 3.3 At one point it even simulated doing the work, log output and everything and when I challenged it about the missing artefacts it actually started questioning my intelligence. Seems appropriate for a Zuck enterprise.
Qwen on the other hand got straight to work with astonishing competency on the same system.
From what I read llama3 needs beefier compute to reliably invoke tools, which I presume relates to it focussing more on simulating AGI rather than being a useful tool.
I have been having pretty good success with Qwen 3.5 9B for "nontrivial but not challenging work all things considered" -- it runs great on my 24gb unified memory m4 pro MacBook Pro. What do the baseline specs look like Mac-wise for getting this model to run? Am I looking at a 96gb? 128? 256?
How does llama.cpp use the GPU efficiently as opposed to MLX?
Is there any way to use MLX and GPU at the same time? Or does memory become a big problem?
TBH, I never understood Apple hyping these neural cores because I didn't think anyone actually uses them except maybe certain photo/video editing software.
If I can generate voice at the same time as video, that would be useful.
None of the examples reflect 'real work', at least not what I'd consider real work. Being able to nail a zero-shot greenfield project is relatively easy even for a small model. There's not much context to build up and it can fall back to similar examples in the training data easily. So long as you're not asking it to invent something wholly new it'll probably manage.
The real test is whether or not it can work with your existing codebases. In my limited experiments Qwen 3.5 (maybe 3.6 is loads better) does OK on a Rust+React app, and less well on a C# monolith. Not to the point of being unusable but definitely poorly enough that I went back to Claude after 20 minutes. If I lost access to a cloud model and had to use Qwen instead I'd be visibly sad.
The question has to be asked: is it (small models struggling at brownfield) an intelligence problem or a context problem? Even if it is an intelligence problem, is it possible to use a customized harness to achieve the same level of performance as SoTA models? That would be very valuable, I think.
none of these local models are good for development, complete waste of time. nobody has $100k+ hardware sitting around at home to actually run a good model
Strix Halo user here. While Qwen 3.6 27B exhibits remarkable intelligence density, I will still take unsloth's dynamic IQ2_XXS of Minimax M2.7 over Q8_0 Qwen 3.6 27B any day of the week, and this isn't just because of generation speed either. I wrote my own custom harness, and I get hallucinated tool call parameters and bizarre invocations with Q3.6 27B even at Q8_0, but no issues with the IQ2_XXS of M2.7.
The article is based on running Qwen 3.6 on a 128GB MacBook Pro. For reference, a 128GB MBP currently starts at $6699 USD [0]
Some people will be happy to pay that premium for privacy, but at roughly 10X the cost of a MacBook Neo, that money could also buy a lot of credits on OpenRouter or frontier labs.
I'm running it on my 4070 12gb with 96gb mem, I'm very happy with the results even if I have to wait a couple minutes for results. To me this is far better than I expected and will continue to use it and improve with skills.md. Pi.dev is amazing by the way.
I was interested to see that Qwen3.5-122B-A10B narrowly beat Qwen3.6-27B on Donato Capitella's SWEBench-verified-mini run with a similar 128GB UMA architecture.
Something I find really confusing from this post is the MLX versions of the model running much slower. As I understand it, these model versions are meant to take advantage of Apple Silicon and MacOS APIs, and should produce better/faster results. Any insight into what’s happening here?
Gemma4 31B with MTP enabled is faster and I feel a bit stronger at coding. Either one can run in 32GB VRAM or unified RAM with some tuning (3 bit weights, 8 bit kv cache)
I've been working with local models for the past year. There's so many possibilities, but I don't think coding is one. Coding requires so many layers beyond inference; I spent so much time trying to replicate what Claude Code does end to end locally. Understanding all the layers and keeping up with the advancements feels like a slog. Even this article messes up and misunderstands what some of the settings are doing. Qwen in particular seems to work at first, then often gets stuck in thought loops when used for actual work.
However, text-to-speech, speech-to-text, and non-code LLM use cases are so useful to have local, and don't require big hardware.
Having a universal reliable inference engine interface, I think, is the big unlock that needs to happen before app devs can ship these features.
Personal concrete use case: meeting recording app. This uses Parakeet + Qwen to create local transcriptions and post-cleanup, respectively.
Right now this app has to download and manage all these models, then bundle an inference engine to run them. It's a lot of code that probably should belong to the OS, or at least a standard interface.
While apps can offload some of this to llama.cpp or a similar process over http, that's another set of setup for the user to do before they can have a useful app.
Anyway, if you're getting started on a Mac, I'd suggest trying out oMLX (https://github.com/jundot/omlx) before messing with llama.cpp. In particular they have community benchmarks so you can see what kind of performance you're likely to get: https://omlx.ai/benchmarks. I wished each one had more configuration details though.
I feel like I'm going insane seeing people buy these 128gb MBP for thousands of dollars to run models that are objectively much worse than SOTA and spending so much more. The amount spent on a 128gb M5 MAX can buy you a damned new car here. What the hell am I missing? Are developers in other countries living in such different worlds?
(I'm aware the price is, in absolute terms, more expensive where I live compared to the USA. That reinforces what I think, because anyone sane that would've bought one of those in another country would sell them as soon as they landed here and save that money.)
Well, I can tell you how my thinking goes: 1) I don't buy my computer just to run LLMs and there are many scenarios where I benefit from both a decent GPU and from a large amount of RAM, 2) I run a solo-founder business which owns exactly one computer in the entire company so it might as well be a good one, and 3) I don't need a new car, so comparing pricing this way is irrelevant.
In other words, yes, buying this kind of machine only to run an LLM locally doesn't make sense, because local LLMs generally still suck for serious programming work (they work great for spam filtering though!). But more generally this machine makes sense for a lot of people.
27-30B in general seems to be the level where you actually start having decent models. I just wish consumer hardware hadn't stagnated so much that we can't easily go higher than that, and that even running those requires a $5k machine now.
137 comments
[ 0.22 ms ] story [ 31.3 ms ] threadQwen on the other hand got straight to work with astonishing competency on the same system.
From what I read llama3 needs beefier compute to reliably invoke tools, which I presume relates to it focussing more on simulating AGI rather than being a useful tool.
Is there any way to use MLX and GPU at the same time? Or does memory become a big problem?
TBH, I never understood Apple hyping these neural cores because I didn't think anyone actually uses them except maybe certain photo/video editing software.
If I can generate voice at the same time as video, that would be useful.
>
> --jinja for tool calling support
Pretty sure this flag hasn't done anything for a while. It's enabled by default since ~November of last year
Local development for who? How many of y'all are rocking 128GB of memory? Am I reading Apple's site correctly that it's a $10,000 laptop?
The real test is whether or not it can work with your existing codebases. In my limited experiments Qwen 3.5 (maybe 3.6 is loads better) does OK on a Rust+React app, and less well on a C# monolith. Not to the point of being unusable but definitely poorly enough that I went back to Claude after 20 minutes. If I lost access to a cloud model and had to use Qwen instead I'd be visibly sad.
Some people will be happy to pay that premium for privacy, but at roughly 10X the cost of a MacBook Neo, that money could also buy a lot of credits on OpenRouter or frontier labs.
[0]: https://www.apple.com/shop/buy-mac/macbook-pro/14-inch-space...
https://pi-local-coding-bench.dev
QAT, MTP, 128k context.
I liked Qwen 3.6 27b too, it just seems that Gemma4 is a bit underrated.
I'm running the NVFP4 alongside Gemma4 at the same quant on an OEM Spark
However, text-to-speech, speech-to-text, and non-code LLM use cases are so useful to have local, and don't require big hardware.
Having a universal reliable inference engine interface, I think, is the big unlock that needs to happen before app devs can ship these features.
Personal concrete use case: meeting recording app. This uses Parakeet + Qwen to create local transcriptions and post-cleanup, respectively.
Right now this app has to download and manage all these models, then bundle an inference engine to run them. It's a lot of code that probably should belong to the OS, or at least a standard interface.
While apps can offload some of this to llama.cpp or a similar process over http, that's another set of setup for the user to do before they can have a useful app.
Anyway, if you're getting started on a Mac, I'd suggest trying out oMLX (https://github.com/jundot/omlx) before messing with llama.cpp. In particular they have community benchmarks so you can see what kind of performance you're likely to get: https://omlx.ai/benchmarks. I wished each one had more configuration details though.
(I'm aware the price is, in absolute terms, more expensive where I live compared to the USA. That reinforces what I think, because anyone sane that would've bought one of those in another country would sell them as soon as they landed here and save that money.)
In other words, yes, buying this kind of machine only to run an LLM locally doesn't make sense, because local LLMs generally still suck for serious programming work (they work great for spam filtering though!). But more generally this machine makes sense for a lot of people.
Jackrong has a few different ones available depending on what you're trying to do: https://huggingface.co/Jackrong