Qwen3.6 LLM runs locally on MacBook Pro via oMLX, achieving fast speeds

A user has successfully run the Qwen3.6-35B-A3B large language model locally on a MacBook Pro using the oMLX framework and openCode. This setup allows for fast token output, estimated at around 100 tokens per second, enabling both chat and coding tasks without an internet connection. The model, a Mixture of Experts type, utilizes approximately 3 billion active parameters per token, contributing to its speed and manageable size after 4-bit quantization, making it suitable for Apple Silicon's unified memory architecture. The setup requires significant RAM, around 20GB, and can cause the GPU to run at 100% capacity during operation. AI

IMPACT Enables local, cost-free LLM usage for chat and coding on consumer hardware, preserving data privacy.

RANK_REASON User guide on running a specific LLM locally using particular software and hardware.

Read on dev.to — LLM tag →

• Apple Silicon
• LM Studio
• MacBook Pro
• Metal
• Mlx
• Ollama
• Omlx Local Ai Models
• openCode
• Open-WebUI
• Qwen3.6 35B-A3B

• model release
• product
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →