How much VRAM does GLM-5 need?

GLM-5 (744B parameters) requires approximately 467.3 GB of memory with Q4_K_M quantization.

What is the best quantization for GLM-5?

The recommended quantization for GLM-5 is Q4_K_M, which balances quality and memory efficiency.

Can it run?

No, GLM-5 requires more memory than RTX 6000 Ada 48GB provides.

FWon't run

Too heavy

Using Q4_K_M in vLLM

Capabilities:

Fit status

Too heavy

Decode

4.4 tok/s

TTFT

43603 ms

Safe context

Memory

467.3 GB / 48.0 GB

Weights453.8 GB

KV Cache6.3 GB

Runtime2.4 GB

Headroom4.8 GB

Workload	Grade	Fit	Decode	TTFT	Context
Agentic Coding	F	Too heavy	5.1 tok/s	55688 ms	4K
Chat	F	Too heavy	5.1 tok/s	20883 ms	4K
Coding	F	Too heavy	4.4 tok/s	43603 ms	4K
RAG	F	Too heavy	5.1 tok/s	69609 ms	4K
Reasoning	F	Too heavy	5.1 tok/s	45246 ms	4K

How GLM-5 (744B params) fits at each quantization level on RTX 6000 Ada 48GB (48.0 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	290.2 GB	Low	F0
Q3_K_S	3	364.6 GB	Low	F0
NVFP4	4