How much VRAM does Qwen3-Coder-Next need?

Qwen3-Coder-Next (80B parameters) requires approximately 57.2 GB of memory with Q4_K_M quantization.

What is the best quantization for Qwen3-Coder-Next?

The recommended quantization for Qwen3-Coder-Next is Q4_K_M, which balances quality and memory efficiency.

Can it run?

Yes, AMD Instinct MI210 64GB can run Qwen3-Coder-Next with a C grade (Tight fit). Expected decode speed: 73.2 tok/s.

CUsable

Tight fit

Using Q4_K_M in Ollama

Capabilities:

Fit status

Tight fit

Decode

73.2 tok/s

TTFT

2644 ms

Safe context

18K

Memory

57.2 GB / 64.0 GB

Weights48.8 GB

KV Cache0.8 GB

Runtime1.2 GB

Headroom6.4 GB

Workload	Grade	Fit	Decode	TTFT	Context
Agentic Coding	C	Tight fit	69.2 tok/s	4071 ms	36K
Chat	C	Tight fit	69.2 tok/s	1527 ms	9K
Coding	C	Tight fit	73.2 tok/s	2644 ms	18K
RAG	C	Tight fit	69.2 tok/s	5089 ms	36K
Reasoning	C	Tight fit	69.2 tok/s	3308 ms	18K

How Qwen3-Coder-Next (80B params) fits at each quantization level on AMD Instinct MI210 64GB (64.0 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	31.2 GB	Low	D40
Q3_K_S	3	39.2 GB	Low	C43
NVFP4	4

Ollama

ollama run qwen-3-coder-next

HuggingFace

huggingface-cli download qwen-3-coder-next

Upgrade options

Mac Studio M2 Ultra 128GBBudget pick

C28.8 tok/s decode

~$3,999 MSRP

NVIDIA A100 80GBBest value

B106.4 tok/s decode

~$15,000 MSRP

NVIDIA A800 80GBBiggest leap

B93.7 tok/s decode