Yes — with 28.3 GB to spare
Mistral Small 4 119B-A6B at Q4_K_M fits your Ryzen AI Max+ 395 · 128 GB entirely in unified memory at 8K context, at an estimated 18 tokens per second. There is room for its full 256K window.
Fully in unified memory
8K context
Q4_K_M · 66.9 GB
Apache 2.0
Released 17 Mar 2026
Vision
Not in the Ollama library
Instruct, reasoning, vision and code in one 119B MoE. A latent KV cache (320 wide) keeps context cheap; the weights still want 64 GB+.
The VRAM budget
weights 66.9 GB
Weights 66.9 GB
KV cache @ 8K 0.18 GB
Runtime overhead 0.6 GB
Free 28.3 GB of 96.0 GB
Quantisation ladder
| Quant | Weights | Total @ 8K | Max context | Tok/s | Quality | Fit |
|---|---|---|---|---|---|---|
| Q8_0 | 117.8 GB | 118.5 GB | — | ~10 | −0.1% ppl | 22.5 GB over |
| Q6_K | 90.9 GB | 91.7 GB | 205K | 13 | −0.4% ppl | Long context |
| Q5_K_M | 78.5 GB | 79.3 GB | 256K | 16 | −0.8% ppl | Long context |
| Q4_K_M | 66.9 GB | 67.7 GB | 256K | 18 | −1.9% ppl | Recommended |
| Q3_K_M | 54.2 GB | 54.9 GB | 256K | 23 | −5.4% ppl | Long context |
| Q2_K | 46.4 GB | 47.2 GB | 256K | 26 | −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 uses multi-head latent attention, so its cache is a compressed latent rather than full K and V.
How to run it
$ llama-server \
-hf mistralai/Mistral-Small-4-119B-2603:Q4_K_M \
-c 8192 -ngl 99
The engine underneath most of the others. Every knob is exposed. More on llama.cpp.
01Download is 66.9 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 96.0 GB of 128 GB. Raising it with iogpu.wired_limit_mb is possible, and risky.
03There is room to go to the model's full 256K context on this card.