shieldstackllc/GLM-4.7-Flash-PRISM-mlx-8bit

<p align="center"> <a href="https://vmlx.net"> <img src="vmlx-logo.png" alt="vMLX" width="120"> </a> </p>

GLM-4.7-Flash-PRISM — MLX 8-bit

MLX 8-bit quantized version of Ex0bit/GLM-4.7-Flash-PRISM for efficient local inference on Apple Silicon.

• —Quantization: 8-bit (8.5 bits per weight, group size 64, affine mode)
• —Architecture: GLM-4 MoE Lite — 47 layers, 64 routed experts, 4 active per token
• —Context: 202K tokens
• —Size: ~30 GB

Usage

from mlx_lm import load, generate

model, tokenizer = load("shieldstackllc/GLM-4.7-Flash-PRISM-mlx-8bit")
response = generate(model, tokenizer, prompt="Hello!", verbose=True)

Or with vMLX for native macOS inference.

About

This model is an abliterated (uncensored) variant of GLM-4.7-Flash, a Mixture-of-Experts language model by Zhipu AI / THUDM. The abliteration was done by Ex0bit as part of the PRISM series. MLX quantization by vMLX.

Also Available

• —GLM-4.7-Flash-PRISM MLX 4-bit (~16 GB)

Made for vMLX

This model was converted and optimized for vMLX — a free, open source macOS native MLX inference engine for Apple Silicon. Download vMLX to run this model locally with zero configuration.

Credits

• —Base model: THUDM/GLM-4 by Zhipu AI
• —Abliteration: Ex0bit/GLM-4.7-Flash-PRISM
• —MLX conversion: vMLX — Run AI locally on Mac. No compromises.

Contact

For questions, issues, or collaboration: admin@vmlx.net