GLM-5 Pruned Q4KM GGUF

Pruned and quantized version of GLM-5 in GGUF format.

Model Details

• Base Model: GLM-5 by Zhipu AI / Z.AI
• Quantization: Q4KM (4-bit, medium quality)
• Pruning: Pruned variant for reduced size
• Format: GGUF (compatible with llama.cpp, ollama, etc.)
• File Size: ~218 GB

Usage

With llama.cpp:

llama-server --model GLM-5-pruned-Q4KM.gguf --n-gpu-layers 999 --ctx-size 8192

Notes

This is a community upload of a pruned + quantized GLM-5 model.
Requires significant RAM/VRAM due to the large MoE architecture.