Mistral on AMD Strix Halo
Can I run Mistral Small 24B on an Ryzen AI Max+ 395 128GB?
Yes. Mistral Small 24B at Q4_K_M uses 15.2 GB of the 96.0 GB available on a Ryzen AI Max+ 395 128GB, and runs at about 9.56 tokens per second. You can push the context to 32K.
Runs comfortably
15.2 GB of 96.0 GB · 16%096 GB
Weights
13.4 GB
KV cache
1.3 GB
Runtime overhead
0.5 GB
Mistral Small 24B at Q4_K_M leaves 80.8 GB spare on a Ryzen AI Max+ 395 128GB. There is room to raise the context length or move up a quantisation level.
Generation9.56tok/s
Prompt processing445tok/s
Max context32Ktokens
KV per 1K tokens0GB
Every quantisation of Mistral Small 24B on a Ryzen AI Max+ 395 128GB
Highlighted row is the highest quality that still fits at 8K context.
| Quantisation | bpw | Weights | Total | Verdict | Max ctx | tok/s |
|---|---|---|---|---|---|---|
| FP16 / BF16 | 16.00 | 43.9 GB | 45.7 GB | Runs comfortably | 32K | 3.02 |
| INT8 / W8A8 | 8.50 | 23.0 GB | 24.8 GB | Runs comfortably | 32K | 5.68 |
| Q8_0 (GGUF) | 8.50 | 23.0 GB | 24.8 GB | Runs comfortably | 32K | 5.68 |
| FP8 (E4M3) | 8.00 | 22.0 GB | 23.7 GB | Runs comfortably | 32K | 5.95 |
| Q6_K | 6.56 | 18.0 GB | 19.8 GB | Runs comfortably | 32K | 7.21 |
| Q5_K_M | 5.67 | 15.6 GB | 17.3 GB | Runs comfortably | 32K | 8.29 |
| Q5_K_S | 5.52 | 15.1 GB | 16.9 GB | Runs comfortably | 32K | 8.51 |
| AWQ 4-bit | 4.25 | 13.5 GB | 15.3 GB | Runs comfortably | 32K | 9.50 |
| GPTQ 4-bit | 4.25 | 13.5 GB | 15.3 GB | Runs comfortably | 32K | 9.50 |
| MXFP4 | 4.25 | 13.5 GB | 15.3 GB | Runs comfortably | 32K | 9.50 |
| Q4_K_M | 4.85 | 13.4 GB | 15.2 GB | Runs comfortably | 32K | 9.56 |
| Q4_K_S | 4.58 | 12.7 GB | 14.5 GB | Runs comfortably | 32K | 10.1 |
| Q4_0 | 4.55 | 12.6 GB | 14.4 GB | Runs comfortably | 32K | 10.1 |
| IQ4_XS | 4.25 | 11.9 GB | 13.6 GB | Runs comfortably | 32K | 10.7 |
| Q3_K_M | 3.91 | 11.0 GB | 12.8 GB | Runs comfortably | 32K | 11.6 |
| IQ3_M | 3.70 | 10.4 GB | 12.2 GB | Runs comfortably | 32K | 12.1 |
| IQ3_XXS | 3.06 | 8.8 GB | 10.6 GB | Runs comfortably | 32K | 14.3 |
| Q2_K | 2.63 | 7.7 GB | 9.5 GB | Runs comfortably | 32K | 16.2 |
| IQ2_XXS | 2.06 | 6.2 GB | 8.0 GB | Runs comfortably | 32K | 19.6 |
| IQ1_M | 1.75 | 5.4 GB | 7.2 GB | Runs comfortably | 32K | 22.2 |