Mistral on Apple M4

Can I run Mistral NeMo 12B on an Mac Mini M4 Pro 48GB?

Yes. Mistral NeMo 12B at Q4_K_M uses 8.8 GB of the 36.0 GB available on a Mac Mini M4 Pro 48GB, and runs at about 22.9 tokens per second. You can push the context to 128K.

Runs comfortably

8.8 GB of 36.0 GB · 24%
036 GB
Weights 7.0 GB
KV cache 1.3 GB
Runtime overhead 0.5 GB

Mistral NeMo 12B at Q4_K_M leaves 27.2 GB spare on a Mac Mini M4 Pro 48GB. There is room to raise the context length or move up a quantisation level.

Generation22.9tok/s
Prompt processing291tok/s
Max context128Ktokens
KV per 1K tokens0GB

Every quantisation of Mistral NeMo 12B on a Mac Mini M4 Pro 48GB

Highlighted row is the highest quality that still fits at 8K context.

QuantisationbpwWeights TotalVerdictMax ctxtok/s
FP16 / BF16 16.00 22.8 GB 24.6 GB Runs comfortably 81K 7.48
INT8 / W8A8 8.50 11.8 GB 13.6 GB Runs comfortably 128K 14.1
Q8_0 (GGUF) 8.50 11.8 GB 13.6 GB Runs comfortably 128K 14.1
FP8 (E4M3) 8.00 11.4 GB 13.2 GB Runs comfortably 128K 14.5
Q6_K 6.56 9.4 GB 11.1 GB Runs comfortably 128K 17.5
Q5_K_M 5.67 8.1 GB 9.9 GB Runs comfortably 128K 20.1
AWQ 4-bit 4.25 7.9 GB 9.7 GB Runs comfortably 128K 20.5
GPTQ 4-bit 4.25 7.9 GB 9.7 GB Runs comfortably 128K 20.5
MXFP4 4.25 7.9 GB 9.7 GB Runs comfortably 128K 20.5
Q5_K_S 5.52 7.9 GB 9.7 GB Runs comfortably 128K 20.6
Q4_K_M 4.85 7.0 GB 8.8 GB Runs comfortably 128K 22.9
Q4_K_S 4.58 6.7 GB 8.5 GB Runs comfortably 128K 23.9
Q4_0 4.55 6.6 GB 8.4 GB Runs comfortably 128K 24.0
IQ4_XS 4.25 6.3 GB 8.0 GB Runs comfortably 128K 25.4
Q3_K_M 3.91 5.8 GB 7.6 GB Runs comfortably 128K 27.1
IQ3_M 3.70 5.6 GB 7.3 GB Runs comfortably 128K 28.2
IQ3_XXS 3.06 4.7 GB 6.5 GB Runs comfortably 128K 32.5
Q2_K 2.63 4.2 GB 6.0 GB Runs comfortably 128K 36.1
IQ2_XXS 2.06 3.5 GB 5.3 GB Runs comfortably 128K 42.4
IQ1_M 1.75 3.1 GB 4.9 GB Runs comfortably 128K 46.9

Also worth checking