Yes — Mistral Small 3.2 runs on Mac M4 Max 64GB
Mistral Small 3.2 (24B) needs 25GB VRAM at Q8 vs your 64GB unified. Estimated ~8–11 tok/s.
Apache-2.0 24B with vision. Q4 ~14 GB — sweet spot for 16 GB cards.
| quant | weights | + KV 8k | fits 64GB? |
|---|---|---|---|
| Q4 | 14GB | ~15GB | ✓ yes |
| Q5 | 17GB | ~18GB | ✓ yes |
| Q8 | 25GB | ~26GB | ✓ yes |
| FP16 | 48GB | ~49GB | ✓ yes |
File size ≠ runtime. KV grows with context — halve to 4k/2k if long chats OOM. Leave 1–2GB free.
Can Mistral Small 3.2 run on Mac M4 Max 64GB?
Mistral Small 3.2 (24B) needs 25GB VRAM at Q8 vs your 64GB unified. Estimated ~8–11 tok/s. Model context 128k. Compare with all models or re-check with your exact specs on the bench.
Unload with ollama stop, cap layers --num-gpu 28, cut context to 2k. See the OOM fixer on the homepage.
Start Q8. Bigger model at Q4 beats smaller at Q8. Only go Q8 if 2GB+ headroom remains.
About ~8–11 tok/s when 100% on GPU. Any CPU offload drops speed 5–20× — check ollama ps.
Yes — Apache-2.0 allows commercial use.