Question 1

Can NVIDIA GH200 96GB run Meta Llama 3 8B Instruct?

Accepted Answer

Yes, NVIDIA GH200 96GB can run Meta Llama 3 8B Instruct with a C grade (Runs well). Expected decode speed: 590.2 tok/s.

Question 2

How much VRAM does Meta Llama 3 8B Instruct need?

Accepted Answer

Meta Llama 3 8B Instruct (8B parameters) requires approximately 16.9 GB of memory with Q4_K_M quantization.

Question 3

What is the best quantization for Meta Llama 3 8B Instruct?

Accepted Answer

The recommended quantization for Meta Llama 3 8B Instruct is Q4_K_M, which balances quality and memory efficiency.

Workload	Grade	Fit	Decode	TTFT	Context
Agentic Coding	C	Runs well	590.2 tok/s	477 ms	169K
Chat	C	Runs well	590.2 tok/s	350 ms	47K
Coding	C	Runs well	590.2 tok/s	350 ms	91K
RAG	C	Runs well	590.2 tok/s	596 ms	169K
Reasoning	C	Runs well	590.2 tok/s	388 ms	91K

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	3.1 GB	Low	D30
Q3_K_S	3	3.9 GB	Low	D30
NVFP4	4

Can NVIDIA GH200 96GB run Meta Llama 3 8B Instruct?

Memory breakdown