Can My Computer Run Local AI?
This tool looks at what your browser can tell about your computer - memory, processor, and graphics card - and gives you a rough idea of which local AI tool and model size would run well on your machine.
Runs entirely in your browser. Nothing about your computer is sent to Reupload or anyone else.
Why Run AI Locally?
Running an AI model locally means it runs entirely on your own computer - no data leaves your machine, no subscription, no usage limits, and it keeps working without an internet connection.
The trade-off is capability: a model that fits on a laptop is smaller than the frontier models behind ChatGPT, Claude, or Gemini, which run on data-center-scale hardware. For everyday tasks - drafting text, summarizing documents, coding help, brainstorming - a well-chosen local model gets remarkably close, especially now that quantization lets even modest hardware run genuinely useful models.
Hardware Tiers at a Glance
Prefer a quick reference to the check above? Here's roughly what each tier of hardware can run.
| Tier | RAM | GPU | Model size | Example |
|---|---|---|---|---|
| Light | 4-8 GB | Integrated / none | 2B-4B | Gemma 2 2B, Phi-4 Mini |
| Medium | 8-16 GB | 6-12 GB VRAM | 7B-14B | Llama 3.2 8B, Mistral 7B |
| High | 16 GB+ | 16 GB+ VRAM | 30B-70B* | Qwen 2.5 32B, Llama 3.1 70B |
* Large models need quantization (a reduced-precision format) to fit on consumer hardware - see the model comparison below for details.
Browse Each Model
Frequently Asked Questions
Do I need a GPU to run AI locally?
No, but it helps a lot. A modern CPU with enough RAM can run small quantized models (2B-8B parameters) at a usable speed; a dedicated GPU with enough VRAM lets you run bigger models much faster.
How much RAM do I need?
At least 8 GB for small models, 16 GB or more for comfortable use of mid-sized models, and 32 GB+ if you want to try large models on CPU alone.
Is running AI locally free?
The software (Ollama, LM Studio, and similar tools) and the open-weight models themselves are free to download. You're paying with your own hardware and electricity instead of a monthly subscription.
Is a local model as good as ChatGPT or Claude?
Not quite on complex reasoning - frontier cloud models still lead there - but for everyday writing, summarizing, and coding help, a well-chosen local model gets close, and plenty of tasks don't need frontier-level capability at all.
Which tool should I use to run a local model?
It depends on your comfort with the command line: Ollama offers a simple CLI, while LM Studio, Jan, GPT4All, and Msty offer a full graphical interface. The check above recommends one based on your hardware.
Can I run local AI on a laptop?
Yes. Many laptops from the last few years - especially Apple Silicon Macs, which share memory between CPU and GPU - run 7B-8B models comfortably.