Yes — with 3.8 GB to spare
Gemma 4 31B at Q4_K_M fits your M1 Pro · 32 GB entirely in unified memory at 8K context, at an estimated 6.4 tokens per second. Past 32K the KV cache pushes it over — quantise the cache to q8_0, or step down a quantisation, to go longer.
Fully in unified memory
8K context
Q4_K_M · 17.6 GB
Apache 2.0
Released 2 Apr 2026
Vision
The dense flagship: strongest maths of the 24–32 GB class (89% AIME), clean prose, vision. Q4 is a tight 24 GB fit.
The VRAM budget
weights 17.6 GB
Weights 17.6 GB
KV cache @ 8K 2.03 GB
Runtime overhead 0.6 GB
Free 3.8 GB of 24.0 GB
Quantisation ladder
| Quant | Weights | Total @ 8K | Max context | Tok/s | Quality | Fit |
|---|---|---|---|---|---|---|
| Q8_0 | 31.0 GB | 33.6 GB | — | ~3.6 | −0.1% ppl | 9.6 GB over |
| Q6_K | 23.9 GB | 26.5 GB | — | ~4.7 | −0.4% ppl | 2.5 GB over |
| Q5_K_M | 20.7 GB | 23.3 GB | 12K | 5.4 | −0.8% ppl | Fits |
| Q4_K_M | 17.6 GB | 20.2 GB | 32K | 6.4 | −1.9% ppl | Recommended |
| Q3_K_M | 14.2 GB | 16.9 GB | 53K | 7.8 | −5.4% ppl | Long context |
| Q2_K | 12.2 GB | 14.8 GB | 66K | 9.2 | −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. This model interleaves sliding-window layers (1024 tokens, 10 of 60 layers global), which is why its cache barely grows with context.
How to run it
$ pip install mlx-lm $ mlx_lm.generate --model mlx-community/gemma-4-31B-it-4bit \ --max-tokens 512 --prompt "Hello"
Apple's own array framework. The fastest path on Apple Silicon. More on MLX.
01Download is 17.6 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 24.0 GB of 32 GB. Raising it with iogpu.wired_limit_mb is possible, and risky.
03There is room to go to 32K context on this card.