What computer hardware a local model needs

Will my laptop or office PC run a local model?

Probably, if it's recent and has enough memory. Both makers' guides give memory figures, which the next section covers.

The makers publish rough guides, and these figures were checked on 28 October 2026 against LM Studio's and Ollama's pages.¹²³ They'll change. Look at the linked pages before you buy anything.

How much memory do the tools suggest?

LM Studio's system requirements page gives different advice for each kind of computer.

Ollama's quickstart uses one model as its example. It says the download is about 7.2 GB and recommends 8 GB of available VRAM, or unified memory on a Mac.² Treat that as one data point, not a rule for every model.

What are RAM, VRAM and unified memory?

RAM is your computer's main working memory. A graphics card (GPU) has its own memory, called VRAM. On a Mac with Apple Silicon, the chip and the graphics share one pool, which Ollama calls unified memory.²

Why care? A model runs fastest when it fits inside the graphics memory. Ollama says that with less VRAM it can use system RAM, but responses may be slower.²

Which graphics cards work?

It depends on the tool. For Nvidia, Ollama's hardware page says compute capability 5.0+ needs driver version 550 and newer, but compute capability 5.0 through 6.2 requires driver version 570 or newer.³ It also says it supports GPU acceleration on Apple devices through the Metal API.³

The same page covers AMD cards, through ROCm and Vulkan.³ For Intel on Linux, it links to driver instructions.³ Check your exact card there before assuming.

No graphics card. Is that a dead end?

Not necessarily.

With less VRAM, Ollama can use system RAM, though it says responses may be slower.²

You can see where a model landed. Ollama's ollama ps command shows what's loaded into memory. Its Processor column says whether that's GPU, CPU or a mix.⁴ For example, 100% CPU means the model was loaded entirely in system memory.⁴

What should I check first on my own machine?

• How much RAM it has.
• Whether it has a separate graphics card, and how much VRAM that card has.
• On a Mac, whether it's Apple Silicon or Intel.
• How much free storage there is, since model files are large. The quickstart's example download alone is about 7.2 GB.²

Next, the sizes page shows how to match a model to what you have. If a laptop is borderline, a smaller file made through quantization is worth a look.