Home / Guides

Guides

Guides

Practical notes for local AI hardware. Published on this site only.

guide

What AI Models Can You Run on a 8GB GPU?

An 8GB GPU can run many 1B–8B language models locally, especially in 4-bit quantization, but context length and runtime overhead matter as much as model size. This guide explains what fits in VRAM, when offload is required, and how to check a specific model before downloading it.

guide

How Much VRAM Do I Need for Local AI?

Estimate the GPU memory needed for local AI by accounting for model weights, quantization, context length, KV cache, and runtime overhead. Use a practical workflow before choosing a GPU.