How much VRAM does Qwen3.5 122B A10B need?

Qwen3.5 122B A10B (122B parameters) requires approximately 93.9 GB of memory with Q3_K_M quantization.

What is the best quantization for Qwen3.5 122B A10B?

The recommended quantization for Qwen3.5 122B A10B is Q3_K_M, which balances quality and memory efficiency.

Can it run?

Yes, Mac Studio M2 Ultra 128GB can run Qwen3.5 122B A10B with a C grade (Runs with offload). Expected decode speed: 8.0 tok/s.

CUsable

Runs with offload

Using Q3_K_M in Ollama

Capabilities:

Fit status

Runs with offload

Decode

8.0 tok/s

TTFT

24063 ms

Safe context

16K

Memory

93.9 GB / 92.2 GB

Weights59.8 GB

KV Cache19.1 GB

Runtime1.2 GB

Headroom13.8 GB

Workload	Grade	Fit	Decode	TTFT	Context
Agentic Coding	F	Too heavy	7.2 tok/s	39011 ms	26K
Chat	C	Tight fit	7.2 tok/s	14629 ms	9K
Coding	C	Runs with offload	8.0 tok/s	24063 ms	16K
RAG	F	Too heavy	7.2 tok/s	48764 ms	26K
Reasoning	C	Runs with offload (needs ~1.1 GB host RAM)	7.2 tok/s	31870 ms	16K

How Qwen3.5 122B A10B (122B params) fits at each quantization level on Mac Studio M2 Ultra 128GB (92.2 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	47.6 GB	Low	C41
Q3_K_S	3	59.8 GB	Low	C43
NVFP4Best for your GPU

HuggingFace

huggingface-cli download hf-unsloth--qwen3-5-122b-a10b-gguf

Upgrade options

AMD Instinct MI350X 288GBBudget pick

C90.9 tok/s decode

~$8,000 MSRP

AMD Instinct MI250X 128GBBest value

C38.8 tok/s decode

~$15,000 MSRP