Unsloth releases quantized versions of GLM-5.2 in GGUF format

🤖 Unsloth has released quantized versions of GLM-5.2 in GGUF format.

A wide spectrum of quantization levels is available on Hugging Face: from BF16 and Q8_0 to extremely compressed IQ1, IQ2, IQ3, and K-Quants. An importance matrix (imatrix) was used to minimize quality loss during heavy compression.

🌍 The availability of optimized GGUF formats accelerates the adoption of cutting-edge architectures into local solutions and lowers the barrier to entry for developers with limited computational resources.

👤 Now, the powerful GLM-5.2 model can be run on a standard home computer or devices with low VRAM, allowing users to choose a balance between speed and quality.

Source 1: https://huggingface.co/unsloth/GLM-5.2-GGUF/tree/main