Explore live open-source projects and AI models.

PMetal: high-performance Apple Silicon framework for local LLM inference, LoRA/QLoRA fine-tuning, serving, quantization, and MLX/Metal acceleration.

JANG — GGUF for MLX. YOU MUST USE JANG_Q RUNTIME. Adaptive Mixed-Precision Quantization + Runtime for Apple Silicon

TurboQuant MLX: Flagship high-performance model quantization for Apple Silicon. Squeeze maximum inference speed and memory efficiency out of your MLX models with zero compromises.

Run GreenBitAI's Quantized LLMs on Apple Devices with MLX

GUF to MLX Converter; LM studio advanced tips; Simple text generation using Apple's MLX framework;Enhanced MLX implementation with transformer architecture;Text generation using the popular Transformers library; Optimized inference with 4-bit quantization(unsloth).

MLX Model Quantization Toolkit - Comprehensive collection of Jupyter notebooks for converting and quantizing large language models using Apple's MLX framework, optimized for Apple Silicon devices.

A high-performance Rust implementation of BitNet neural networks featuring revolutionary 1.58-bit quantization, advanced memory management, comprehensive GPU acceleration (Metal + MLX), and production-ready infrastructure optimized for Apple Silicon and beyond.

Utilities to evaluate MLX quantizations

PLE-safe MLX quantization for Google Gemma 4 (E2B/E4B/26B/31B)