Can I Run gemma 4 31B on a Apple M5 Ultra (256GB)?
Runs at full precision (fp16). Zero quality loss.
11 quantizations fit your 256GB
| Quant | Min VRAM | Recommended | File size | Headroom |
|---|---|---|---|---|
| fp16BEST | 66.4 GB | 67.9 GB | 1.2 GB | +189.6 GB |
| Q8_0 | 35.7 GB | 37.2 GB | 32.6 GB | +220.3 GB |
| Q6_K | 27.9 GB | 29.4 GB | 25.2 GB | +228.1 GB |
| Q5_K_M | 24.2 GB | 25.7 GB | 21.9 GB | +231.8 GB |
| Q5_K_S | 23.6 GB | 25.1 GB | 21.3 GB | +232.4 GB |
| Q4_K_M | 20.8 GB | 22.3 GB | 18.7 GB | +235.2 GB |
| Q4_K_S | 19.7 GB | 21.2 GB | 17.8 GB | +236.3 GB |
| Q3_K_L | 15.6 GB | 17.1 GB | 16.6 GB | +240.4 GB |
| Q3_K_M | 14.7 GB | 16.2 GB | 15.3 GB | +241.3 GB |
| Q3_K_S | 13.6 GB | 15.1 GB | 13.8 GB | +242.4 GB |
| Q2_K | 11.8 GB | 13.3 GB | 11.9 GB | +244.3 GB |
Try it in the cloud first
Don't want to download gemma 4 31B just to try it? Use a hosted API or rent a GPU by the second.
Affiliate links — we earn a commission at no cost to you.
All quant variants, benchmark scores, and use-case tags.
Top-ranked open-source models that fit in 256GB.
FAQ
Can the Apple M5 Ultra (256GB) run gemma 4 31B?
Yes. The Apple M5 Ultra (256GB)'s 256GB of unified memory is enough to run gemma 4 31B at fp16 quantization (66.4GB required).
What's the best quantization to use?
fp16 is the highest-precision quantization that fits in your 256GB. It uses about 66.4GB of memory and 67.9GB recommended for comfortable inference.
What if I need more headroom for context length?
KV cache memory grows with context length. The numbers above assume a baseline 2K-4K context. For long-context use (32K+), add another 2-6GB depending on the model architecture.