Apple Silicon Model Catalog

Access our curated collection of quantized architectures on Hugging Face:

• Official Collection: MLX Models by SirSahOl — Optimized for Apple Silicon
• Profile Repositories: SirSahOl on Hugging Face

---

1. Flagship Releases

Model Architecture	Quantization	RAM Footprint	Target Workload	Hugging Face Repository
DeepSeek-R1-Distill-Qwen-14B	4-bit AWQ	~9.2 GB	Heavy Mathematical & Code Reasoning	SirSahOl/DeepSeek-R1-Distill-Qwen-14B-mlx-4bit
Qwen 2.5 7B Instruct	8-bit MLX	~7.8 GB	General Agentic Workflows & Tool Use	SirSahOl/Qwen2.5-7B-Instruct-mlx-8bit
GLM-Edge-4B-Chat	4-bit AWQ	~2.6 GB	Low-Memory Mobile / Base M1 Macs	SirSahOl/glm-edge-4b-chat-mlx-4bit
Llama-Guard-3-1B	8-bit MLX	~1.3 GB	In-Flight Safety & Content Classification	SirSahOl/Llama-Guard-3-1B-mlx-8bit
SmolLM2-135M-Chat	8-bit MLX	~210 MB	Real-Time On-Device Text Autocomplete	SirSahOl/SmolLM2-135M-chat-mlx-8bit

---

2. Quick CLI Pull

To run any model locally with mlx-lm:

pip install mlx-lm
mlx_lm.generate --model SirSahOl/DeepSeek-R1-Distill-Qwen-14B-mlx-4bit \
  --prompt "Write a concurrent Go worker pool."