Find out what your machine can actually run
Local AI models are getting good fast, but most guides assume you have a server rack. This site is built for the machine you already own: a gaming PC, a laptop with a decent GPU, or a Mac. Punch in your hardware and get a straight answer, not a benchmark leaderboard. Once a model's downloaded, it keeps working with no internet connection at all.
Check your hardware first
Before you download anything, find out if it will actually run. The calculator checks your RAM, VRAM, and quantization choice against real model sizes.
Open the hardware calculatorGuides
Local LLM Hardware Guide
GPUs, RAM, Apple Silicon, and what to check when buying used.
Read the guide →Best Local LLM by Hardware Tier
Organized by what fits your machine, not ranked benchmark scores.
Read the guide →How to Run an LLM Locally
The shortest real path from "I have a computer" to talking to a local model.
Read the guide →Best Local LLMs for Coding
Coding-focused models by hardware tier, plus agentic vs. autocomplete.
Read the guide →Apple Silicon Local AI Guide
The current Mac lineup for local AI, including the just-launched M5 Mac Studio.
Read the guide →Run DeepSeek Locally
Real hardware numbers by model size, and why other guides disagree so much.
Read the guide →Run Qwen Locally
Picking the right version, plus two setup gotchas that trip people up.
Read the guide →Browse models
Filter by hardware tier and use case in the model directory, or jump straight to the calculator if you already know what you're checking.