Question 1

Can RTX 3090 Ti 24GB run internlm2 5 20b chat?

Accepted Answer

Yes, RTX 3090 Ti 24GB can run internlm2 5 20b chat with a B grade (Runs well). Expected decode speed: 59.5 tok/s.

Question 2

How much VRAM does internlm2 5 20b chat need?

Accepted Answer

internlm2 5 20b chat (20B parameters) requires approximately 18.9 GB of memory with Q4_K_M quantization.

Question 3

What is the best quantization for internlm2 5 20b chat?

Accepted Answer

The recommended quantization for internlm2 5 20b chat is Q4_K_M, which balances quality and memory efficiency.

Workload	Grade	Fit	Decode	TTFT	Context
Agentic Coding	C	Tight fit	58.7 tok/s	4800 ms	35K
Chat	B	Runs well	58.7 tok/s	1800 ms	11K
Coding	B	Runs well	59.5 tok/s	3254 ms	20K
RAG	C	Tight fit	58.7 tok/s	6000 ms	35K
Reasoning	B	Runs well	58.7 tok/s	3900 ms	20K

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	7.8 GB	Low	D36
Q3_K_S	3	9.8 GB	Low	D38
NVFP4	4

Can RTX 3090 Ti 24GB run internlm2 5 20b chat?

Memory breakdown