How much VRAM does Qwen 3 235B A22B need?

Qwen 3 235B A22B (235B parameters) requires approximately 175.3 GB of memory with Q4_K_M quantization.

What is the best quantization for Qwen 3 235B A22B?

The recommended quantization for Qwen 3 235B A22B is Q4_K_M, which balances quality and memory efficiency.

Can it run?

Yes, Mac Studio M3 Ultra 256GB can run Qwen 3 235B A22B with a C grade (Runs with offload). Expected decode speed: 9.9 tok/s.

CUsable

Runs with offload

Using Q4_K_M in llama.cpp

Capabilities:

Fit status

Runs with offload

Decode

9.9 tok/s

TTFT

19544 ms

Safe context

17K

Memory

175.3 GB / 184.3 GB

Weights143.4 GB

KV Cache3.4 GB

Runtime0.9 GB

Headroom27.6 GB

Workload	Grade	Fit	Decode	TTFT	Context
Agentic Coding	C	Runs with offload	9.9 tok/s	28428 ms	33K
Chat	C	Tight fit	9.9 tok/s	10660 ms	8K
Coding	C	Runs with offload	9.9 tok/s	19544 ms	17K
RAG	C	Runs with offload	9.9 tok/s	35535 ms	33K
Reasoning	C	Runs with offload	9.9 tok/s	23097 ms	17K

How Qwen 3 235B A22B (235B params) fits at each quantization level on Mac Studio M3 Ultra 256GB (184.3 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	91.7 GB	Low	C40
Q3_K_S	3	115.2 GB	Low	C43
NVFP4	4

HuggingFace

huggingface-cli download qwen-3-235b-a22b

Upgrade options

AMD Instinct MI350X 288GBBudget pick

B108.7 tok/s decode

~$8,000 MSRP

AMD Instinct MI300X 192GBBest value

C76.9 tok/s decode

~$15,000 MSRP