Knowledge Base · Local AI

Can my laptop run a local AI model?

Updated October 2026. No affiliate links on this page. For most people the answer is yes. Here's what your laptop can handle and the one number that decides it.

If your laptop has at least 8 GB of RAM, it can run a local AI model today. How far you can go depends on memory:

8 GB: small models, and they run well.
16 GB: 7B and 8B models, which are useful for real work.
32 GB: 13B and 14B models.
64 GB or more: large models.

Memory is the spec that matters. Processor speed matters far less.

Words you'll see in this guide

LLM: large language model, the kind of AI behind ChatGPT and Claude. A local LLM runs on your own computer.
RAM: your computer's main working memory. More RAM lets it run more apps at once. AI models can run from it too, just slower than from VRAM.
VRAM: the memory on your graphics card. AI models load into it, so more VRAM means you can run bigger models.
Quantization: shrinking a model by storing its numbers less precisely, so it fits in less memory with a small drop in quality. Think of saving a photo as a smaller JPEG. "4-bit" (Q4) is the common setting.
Parameters (3B, 8B, 70B): a rough measure of a model's size. "8B" means 8 billion. Bigger models are usually smarter and need more memory.

See the full glossary

Why memory decides everything

When an AI model runs, it has to fit into memory: either your system RAM, where the processor runs it, or a graphics card's VRAM, where the card runs it. If it fits, it runs. If it doesn't, it won't load at all. That's why memory, not processor speed, is what limits you.

Models are almost always run in a compressed form called quantized, which uses about a quarter of the memory of the full version.

What runs at each amount of RAM

Your RAM	What you can run	What it's like
8 GB	Small models (1B to 3B), like Llama 3.2 3B or Phi	Fine for simple chat, summaries, and experimenting
16 GB	7B and 8B models, like Llama 3.1 8B or Mistral	The sweet spot. Useful for real work, on par with older cloud models
32 GB	13B and 14B models	Noticeably smarter answers, still comfortable
64 GB or more	Large models, up to 70B on a Mac	Getting close to cloud quality, especially on Apple Silicon

The math behind the table: a compressed model needs about 0.5 to 1 GB of memory per billion parameters, plus a couple of GB extra.

Macs get more out of their memory. Apple Silicon laptops share one pool of memory between the processor and graphics, so their RAM works like VRAM. A 32 GB MacBook runs models that a 32 GB Windows laptop without a big graphics card can't. If you have an M-series Mac, go one row further down the table. More in Mac vs PC for local AI.

Fitting vs running fast

Fitting the model and running it fast are two different things. A model running on a graphics card, or in a Mac's shared memory, writes quickly. One that only fits in regular system RAM runs on the processor. You get the same answers, several times slower.

That's why a gaming laptop with an Nvidia graphics card feels much faster than a thin laptop with the same RAM: the model lives in the card's fast VRAM. Either way, if it fits, it works.

How to try it

Start with an app that hides the technical parts.

LM Studio is the friendliest first try. It's a normal app: you browse and download models in a couple of clicks, and it tells you whether each one fits your laptop.
Ollama is the favorite if you're comfortable typing commands.

Not sure which suits you? See Ollama vs LM Studio.

Check your exact laptop

Type your chip and RAM, or your graphics card, into the Server Calculator. It tells you what fits and roughly how fast each model size will run.

Thinking about an upgrade? See how much RAM you need for a local LLM.
Thinking about a dedicated machine? See the best GPU for local AI under $500.

Quick answers