Llama 3.2 3B Instruct on Apple M3 Pro (18 GB)

Can the Apple M3 Pro (18 GB) run Llama 3.2 3B Instruct?

Yes, it runs great

Runs great. At Q5_K_M it needs about 4.8 GB of 12.1 GB usable, with an estimated speed of 40.8 tok/s.

Best quant: Q5_K_M
Generation speed: 40.8 tok/s
Prompt processing: 1,429 tok/s
First token (1K prompt): 0.7 s
Largest safe context: 69K

Best app to run it on the Apple M3 Pro (18 GB)

On macOS

LM Studio

Desktop app · runs on llama.cpp, MLX · use the MLX build

Runs the MLX build, typically 35–90% faster than GGUF for a model with 3.2B active parameters on Apple silicon.
Point-and-click: search, download and chat in one window.

Open Llama 3.2 3B Instruct in LM Studio

MLX LM Runs the MLX build, typically 35–90% faster than GGUF for a model with 3.2B active parameters on Apple silicon.
Ollama One command with Ollama’s own build of this model.

How we choose: the files published for this model, your operating system and GPU, whether it fits fully in memory, and published speed tests (MLX vs llama.cpp on Apple silicon, vLLM vs llama.cpp vs Ollama on NVIDIA). Details.

Download Llama 3.2 3B Instruct

Recommended for the Apple M3 Pro (18 GB): Q5_K_M. Every other quantization has its own download link in the table below.

Quickest · Ollama

One command downloads a ready-to-run version and starts a chat.

Get Ollama at ollama.com/download (opens in a new tab) · model page (opens in a new tab)

ollama run llama3.2:3b-instruct-q5_K_M

Quantized files · GGUF

Works in LM Studio, llama.cpp, Jan and Ollama. No sign-in needed.

Download the Q5_K_M file (opens in a new tab)

huggingface.co/unsloth/Llama-3.2-3B-Instruct-GGUF (opens in a new tab)

ollama run hf.co/unsloth/Llama-3.2-3B-Instruct-GGUF:Q5_K_M

Original weights · Hugging Face

Full-precision files from Meta, for fine-tuning, vLLM, Transformers.

huggingface.co/meta-llama/Llama-3.2-3B-Instruct (opens in a new tab)

Free Hugging Face account needed: request access on the model page, and Meta approves it (usually within minutes to a few hours).

Official repo · GitHub

Reference code, inference scripts and docs from Meta. The weights themselves download from Hugging Face.

github.com/meta-llama/llama-models (opens in a new tab)

Meta’s official Llama repo: model cards, prompt formats and download tools.

git clone https://github.com/meta-llama/llama-models

Links checked 3 days ago against Hugging Face, Ollama and GitHub.

How it’s rated

Quality 33Basic quality (D) · 33 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.: How capable the model itself is at Q5_K_M, from its editorial quality tier scaled by the quantization or precision. It’s the same on any hardware that runs this variant. Scale: A+ 85+ · A 70+ · B 55+ · C 40+ · D under 40.
Runs Runs great: How well it runs on the Apple M3 Pro (18 GB): speed, capped by how well it fits in memory. Runs great: 30+ tokens/s (or near the image, video or decision speed target) · Runs well: 10+ tokens/s · Slow: 3+ tokens/s · Very slow: under 3 · Won’t run: doesn’t fit even with offload. A tight fit or RAM offload caps it at Runs well; heavy offload caps it at Slow.

Match score

To order lists, each result also gets a match score out of 100 from four parts. It favors the most capable model that still runs well.

Model qualityEditorial quality tier of the model, scaled by the quantization or precision used.	15.3 / 45
Memory fitFull marks when the model fits in the share of unified memory the GPU can use; fewer when it goes past that limit.	25.0 / 25
SpeedGeneration speed against a 40 tokens/s target for LLMs, or render time against a target for image and video.	15.0 / 15
Memory useRewards using the memory you have, so a bigger, better variant beats a tiny one that leaves most memory idle.	9.2 / 15
Match score	64.4 / 100

The match score orders results; it is not shown as a grade. Quality and Runs are the two ratings in the tables.

Every quantization on the Apple M3 Pro (18 GB)

Quant	Memory needed	Available	Speed	Runs	Quality	Download
`Q2_K`	3.8 GB	12.1 GB	81.1 tok/s	Runs great	24Basic quality (D) · 24 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 58/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q3_K_M`	4.1 GB	12.1 GB	63.9 tok/s	Runs great	29Basic quality (D) · 29 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q4_K_M`	4.5 GB	12.1 GB	49.4 tok/s	Runs great	33Basic quality (D) · 33 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q5_K_M`	4.8 GB	12.1 GB	40.8 tok/s	Runs great	33Basic quality (D) · 33 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q6_K`	5.1 GB	12.1 GB	34.3 tok/s	Runs great	34Basic quality (D) · 34 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`Q8_0`	5.9 GB	12.1 GB	25.1 tok/s	Runs well	35Basic quality (D) · 35 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab)
`F16`	8.9 GB	12.1 GB	11.6 tok/s	Runs well	35Basic quality (D) · 35 of 100: how capable the model is at F16, whatever it runs on. Match score 63/100 (quality, fit, speed and memory use) sets the ranking.	GGUF (opens in a new tab) Original weights (opens in a new tab)

Memory figures assume 8K of context. Longer conversations or documents grow the KV cache.

Other hardware that runs it

35Basic quality (D) · 35 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 4060 Q8_0 · 53.0 tok/s · Runs great
35Basic quality (D) · 35 of 100: how capable the model is at F16, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 4070 SUPER F16 · 45.2 tok/s · Runs great
35Basic quality (D) · 35 of 100: how capable the model is at F16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking. AMD Radeon RX 9070 XT F16 · 49.4 tok/s · Runs great
35Basic quality (D) · 35 of 100: how capable the model is at F16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 4070 Ti SUPER F16 · 60.3 tok/s · Runs great
35Basic quality (D) · 35 of 100: how capable the model is at F16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 4080 SUPER F16 · 66.0 tok/s · Runs great
35Basic quality (D) · 35 of 100: how capable the model is at F16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GeForce RTX 5060 Ti 16GB F16 · 40.2 tok/s · Runs great

Quick answers

How much memory does Llama 3.2 3B Instruct need on the M3 Pro 18GB?

At Q5_K_M it needs about 4.8 GB. The M3 Pro 18GB has 12.1 GB usable.

How fast is Llama 3.2 3B Instruct on the M3 Pro 18GB?

About 40.8 tok/s at Q5_K_M, an estimate from the hardware’s memory bandwidth and the size of the model.

Which quantization of Llama 3.2 3B Instruct should I use on the M3 Pro 18GB?

Q5_K_M. Of the versions that fit, it is the one that runs best; when two run equally well, the check prefers the higher-rated one.

Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.