Quantized GLM MoE model (W4A16 AWQ) trends on Hugging Face

AikidoSec · hf · 2026-09-22

AikidoSec's altar-1 is trending on Hugging Face as a text-generation model built on the GLM MoE architecture (tags include glmmoedsa and glm-5.3), quantized to AWQ INT4 (W4A16) in compressed-tensors format.

The quantization lets large MoE models run with much lower memory footprints, useful for local or self-hosted inference setups.

Original post →

More from Infra

Infra channel →