How Much VRAM Do You Actually Need to Run LLMs Locally?
A practical VRAM guide by model size and quantization — what really fits on 4, 8, 12, 16 and 24 GB cards, tested on real hardware.
// local inference, on a budget
Run AI on your own hardware.
Tested guides for running LLMs locally — on the GPU you have, not the one in the press release. Real setups, real VRAM numbers, no hype.
Latest guides
Run Your First Local LLM with Ollama in 10 Minutes
From zero to chatting with a local model: install Ollama, pick a model that fits your VRAM, and verify it's actually using your GPU.