Question 1

Can NVIDIA GH200 96GB run Qwen2.5 3B Instruct?

Accepted Answer

Yes, NVIDIA GH200 96GB can run Qwen2.5 3B Instruct with a C grade (Runs well). Expected decode speed: 1573.8 tok/s.

Question 2

How much VRAM does Qwen2.5 3B Instruct need?

Accepted Answer

Qwen2.5 3B Instruct (3B parameters) requires approximately 13.4 GB of memory with Q4_K_M quantization.

Question 3

What is the best quantization for Qwen2.5 3B Instruct?

Accepted Answer

The recommended quantization for Qwen2.5 3B Instruct is Q4_K_M, which balances quality and memory efficiency.

Workload	Grade	Fit	Decode	TTFT	Context
Agentic Coding	C	Runs well	1573.8 tok/s	350 ms	226K
Chat	C	Runs well	1573.8 tok/s	350 ms	57K
Coding	C	Runs well	1573.8 tok/s	350 ms	114K
RAG	C	Runs well	1573.8 tok/s	350 ms	226K
Reasoning	C	Runs well	1573.8 tok/s	350 ms	114K

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	1.2 GB	Low	D30
Q3_K_S	3	1.5 GB	Low	D30
NVFP4	4

Can NVIDIA GH200 96GB run Qwen2.5 3B Instruct?

Memory breakdown