Mistral on AMD RDNA3

Can I run Mistral NeMo 12B on an Radeon RX 7900 XTX?

Yes. Mistral NeMo 12B at Q4_K_M uses 9.1 GB of the 21.8 GB available on a Radeon RX 7900 XTX, and runs at about 70.3 tokens per second. You can push the context to 89K.

Runs comfortably

9.1 GB of 21.8 GB · 42%
022 GB
Weights 7.0 GB
KV cache 1.3 GB
Runtime overhead 0.8 GB

Mistral NeMo 12B at Q4_K_M leaves 12.7 GB spare on a Radeon RX 7900 XTX. There is room to raise the context length or move up a quantisation level.

Generation70.3tok/s
Prompt processing2109tok/s
Max context89Ktokens
KV per 1K tokens0GB

Every quantisation of Mistral NeMo 12B on a Radeon RX 7900 XTX

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

QuantisationbpwWeights TotalVerdictMax ctxtok/s
FP16 / BF16 16.00 22.8 GB 24.9 GB Won't fit — 7.49
INT8 / W8A8 8.50 11.8 GB 13.9 GB Runs comfortably 58K 43.4
Q8_0 (GGUF) 8.50 11.8 GB 13.9 GB Runs comfortably 58K 43.4
FP8 (E4M3) 8.00 11.4 GB 13.5 GB Runs comfortably 61K 44.9
Q6_K 6.56 9.4 GB 11.4 GB Runs comfortably 74K 54.0
Q5_K_M 5.67 8.1 GB 10.2 GB Runs comfortably 82K 61.8
AWQ 4-bit 4.25 7.9 GB 10.0 GB Runs comfortably 83K 63.1
GPTQ 4-bit 4.25 7.9 GB 10.0 GB Runs comfortably 83K 63.1
MXFP4 4.25 7.9 GB 10.0 GB Runs comfortably 83K 63.1
Q5_K_S 5.52 7.9 GB 10.0 GB Runs comfortably 84K 63.3
Q4_K_M 4.85 7.0 GB 9.1 GB Runs comfortably 89K 70.3
Q4_K_S 4.58 6.7 GB 8.8 GB Runs comfortably 91K 73.5
Q4_0 4.55 6.6 GB 8.7 GB Runs comfortably 91K 73.9
IQ4_XS 4.25 6.3 GB 8.3 GB Runs comfortably 94K 77.9
Q3_K_M 3.91 5.8 GB 7.9 GB Runs comfortably 97K 83.0
IQ3_M 3.70 5.6 GB 7.6 GB Runs comfortably 98K 86.5
IQ3_XXS 3.06 4.7 GB 6.8 GB Runs comfortably 104K 99.3
Q2_K 2.63 4.2 GB 6.3 GB Runs comfortably 107K 110
IQ2_XXS 2.06 3.5 GB 5.6 GB Runs comfortably 112K 129
IQ1_M 1.75 3.1 GB 5.2 GB Runs comfortably 114K 142

Also worth checking

Mistral NeMo 12B on every GPU
Full hardware tableEverything that runs on a Radeon RX 7900 XTX
Full model tableWhy the KV cache is the part that bites
Guide