No — not on this device
Qwen3 235B-A22B at Q4_K_M needs 134.2 GB against 48.0 GB usable, and the shortfall of 86.2 GB is more than 64 GB of system RAM can cover at a tolerable speed. A smaller sibling or a lower quantisation is the honest answer here.
Does not fit
8K context
Q4_K_M · 132.1 GB
Apache 2.0
Released Apr 2025
Workstation class. Realistically a 192 GB unified-memory or multi-GPU model.
What hardware do I need for Qwen3 235B-A22B? →
Fits instead: Qwen3 32B (21.0 GB) · Qwen3 Coder 30B-A3B (18.5 GB)
The VRAM budget
weights 132.1 GB
Weights 132.1 GB
KV cache @ 8K 1.47 GB
Runtime overhead 0.6 GB
Over budget 86.2 GB past 48.0 GB
Quantisation ladder
| Quant | Weights | Total @ 8K | Max context | Tok/s | Quality | Fit |
|---|---|---|---|---|---|---|
| Q8_0 | 232.5 GB | 234.6 GB | — | ~3.1 | −0.1% ppl | 186.6 GB over |
| Q6_K | 179.5 GB | 181.5 GB | — | ~4.0 | −0.4% ppl | 133.5 GB over |
| Q5_K_M | 155.1 GB | 157.2 GB | — | ~4.6 | −0.8% ppl | 109.2 GB over |
| Q4_K_M | 132.1 GB | 134.2 GB | — | ~5.4 | −1.9% ppl | 86.2 GB over |
| Q3_K_M | 107.0 GB | 109.0 GB | — | ~6.7 | −5.4% ppl | 61.0 GB over |
| Q2_K | 91.6 GB | 93.7 GB | — | ~7.8 | −15% ppl | 45.7 GB over |
Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it.
How to run it
$ llama-server \
-hf Qwen/Qwen3-235B-A22B:Q4_K_M \
-c 8192 -ngl 32
The engine underneath most of the others. Every knob is exposed. More on llama.cpp.
01Download is 132.1 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 48.0 GB of 64 GB. Raising it with iogpu.wired_limit_mb is possible, and risky.
03If it falls back to CPU silently, drop the context first, then step down a quantisation.