Question 1

How much VRAM does GLM-5 need?

Accepted Answer

GLM-5 (753.9B) requires approximately 1685.0176580429077 GB of VRAM at full precision (FP16), 842.5 GB at 8-bit quantization (Q8), and 421.3 GB at 4-bit quantization (Q4). The actual VRAM usage may be slightly higher due to KV cache and runtime overhead.

Question 2

What GPUs can run GLM-5?

Accepted Answer

We currently have no compatible GPUs indexed for GLM-5. This may be because the model's VRAM requirements exceed available GPUs in our database, or VRAM data hasn't been computed yet.

Question 3

What is the cheapest GPU to run GLM-5?

Accepted Answer

We don't currently have cloud pricing data for GPUs compatible with GLM-5. Cloud pricing data is updated regularly — check back soon.

Question 4

Can GLM-5 run on consumer hardware?

Accepted Answer

GLM-5 requires approximately 1685.0176580429077 GB of VRAM at FP16, and even with 4-bit quantization (Q4) it needs about 421.3 GB. This exceeds the 24 GB available on the highest-end consumer GPU (RTX 4090), so GLM-5 requires datacenter or cloud GPUs.

Question 5

What quantization options are available for GLM-5?

Accepted Answer

GLM-5 supports the following quantization levels: FP16 (full precision): ~1685.0176580429077 GB, Q8 (8-bit): ~842.5 GB, Q6 (6-bit): ~631.9 GB, Q4 (4-bit): ~421.3 GB. Lower quantization reduces VRAM usage at the cost of some quality loss. Q8 offers a good balance between quality and memory savings, while Q4 provides maximum compression. As a Mixture of Experts (MoE) model with 256 experts (8 active), GLM-5 has a larger total weight size but only activates a subset of parameters per token, which can affect effective memory usage during inference.

GLM-5

VRAM Requirements

Compatible GPUs

GLM-5 FAQ

VRAM requirements for GLM-5 at different quantization levels
Quantization	VRAM Required
FP16	1685 GB
Q8	842.5 GB
Q6	631.9 GB
Q4	421.3 GB