Yes — with 90.7 GB to spare

Qwen2.5-Coder 7B at Q4_K_M fits your M4 Max · 128 GB entirely in unified memory at 8K context, at an estimated 71 tokens per second. There is room for its full 128K window.

Fully in unified memory 8K context Q4_K_M · 4.3 GB Apache 2.0 Released Nov 2024

The standard local autocomplete model — small enough to keep resident all day, and still the best FIM model under 8B.

What hardware do I need for Qwen2.5-Coder 7B? →

The VRAM budget

weights 4.3 GB

Weights 4.3 GB KV cache @ 8K 0.44 GB Runtime overhead 0.6 GB Free 90.7 GB of 96.0 GB

Quantisation ladder

Quant	Weights	Total @ 8K	Max context	Tok/s	Quality	Fit
F16	14.2 GB	15.2 GB	128K	21	Reference	Long context
Q8_0	7.5 GB	8.6 GB	128K	40	−0.1% ppl	Long context
Q6_K	5.8 GB	6.9 GB	128K	52	−0.4% ppl	Long context
Q5_K_M	5.0 GB	6.1 GB	128K	61	−0.8% ppl	Long context
Q4_K_M	4.3 GB	5.3 GB	128K	71	−1.9% ppl	Recommended
Q3_K_M	3.5 GB	4.5 GB	128K	88	−5.4% ppl	Long context
Q2_K	3.0 GB	4.0 GB	128K	103	−15% ppl	Long context

Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it.

How to run it

Ollama llama.cpp LM Studio MLX

terminal

$ pip install mlx-lm
$ mlx_lm.generate --model mlx-community/Qwen2.5-Coder-7B-Instruct-4bit \
    --max-tokens 512 --prompt "Hello"

Apple's own array framework. The fastest path on Apple Silicon. More on MLX.

01Download is 4.3 GB. Keep it on an SSD — a first load off a spinning disk takes minutes.

02macOS caps what the GPU may wire down at about 96.0 GB of 128 GB. Raising it with iogpu.wired_limit_mb is possible, and risky.

03There is room to go to the model's full 128K context on this card.

See all models for this rig Compare with another model