can i run · seo
Yes — Qwen3 32B runs on Mac M4 Pro 24GB
Qwen3 32B (32B) needs 16GB VRAM at Q3 vs your 24GB unified. Estimated ~6–8 tok/s.
Dense 32B for 24 GB cards at Q4. Serious coding and long-horizon agent work.
best fit on this rig · weights + KV + headroom
Q3 · 16GB weights+1GB KV @ 8kfits GPU~6–8 tok/s
| quant | weights | + KV 8k | fits 24GB? |
|---|---|---|---|
| Q3 | 16GB | ~17GB | ✓ yes |
| Q4 | 20GB | ~21GB | ✓ yes |
| Q5 | 24GB | ~25GB | × tight/no |
| Q8 | 35GB | ~36GB | × tight/no |
| FP16 | 64GB | ~65GB | × tight/no |
File size ≠ runtime. KV grows with context — halve to 4k/2k if long chats OOM. Leave 1–2GB free.
ollama run qwen3:32b
Run it: 1) install Ollama · 2) copy command · 3) verify with ollama ps (want 100% GPU).
faq · straight answers
Can Qwen3 32B run on Mac M4 Pro 24GB?
Qwen3 32B (32B) needs 16GB VRAM at Q3 vs your 24GB unified. Estimated ~6–8 tok/s. Model context 128k. Compare with all models or re-check with your exact specs on the bench.
What if it OOMs?
Unload with ollama stop, cap layers --num-gpu 28, cut context to 2k. See the OOM fixer on the homepage.
Q4 or Q8 on this rig?
Start Q3. Bigger model at Q4 beats smaller at Q8. Only go Q8 if 2GB+ headroom remains.
How fast will it be?
About ~6–8 tok/s when 100% on GPU. Any CPU offload drops speed 5–20× — check ollama ps.
Is Apache-2.0 commercial-safe?
Yes — Apache-2.0 allows commercial use.
also fits Mac M4 Pro 24GB
Qwen3 32B on other rigs