Run Any Open AI Model on Your Own Computer (Ollama & LM Studio)

Check your computer

• RAM is the key number: 8 GB runs small models, 16 GB runs good mid-size models, 32 GB+ runs larger ones.
• A graphics card (NVIDIA with plenty of VRAM) or an Apple Silicon Mac (M1 or newer) makes models much faster.
• Free disk space: 2–10 GB per model for typical sizes.

Path A — Ollama (simple, command-based)

1. 1

   Install

   Download Ollama from ollama.com for your operating system and install it.

2. 2

   Run a model

   Open a terminal and type a command such as ollama run qwen3:4b , ollama run deepseek-r1:7b , ollama run llama3.2 or ollama run gemma3 .

3. 3

   Manage models

   ollama list shows installed models, ollama rm <name> deletes one.

4. 4

   Use it from other apps

   Ollama runs a local server (by default on port 11434) that many chat interfaces, code editors and scripts can connect to.

Path B — LM Studio (everything by clicking)

1. 1

   Install

   Download LM Studio from lmstudio.ai and install it.

2. 2

   Find a model

   Search in the Discover tab; LM Studio shows whether each download fits your memory.

3. 3

   Chat

   Load the model in the Chat tab and start typing.

Add a nicer chat interface (optional)

If you use Ollama, you can add a ChatGPT-style interface with a free project such as Open WebUI. It connects to Ollama automatically and gives you chat history, file uploads and model switching in your browser.

Which model for which job?

• General chat and writing: Qwen, Llama or Gemma in a size that fits your RAM.
• Maths, logic and step-by-step reasoning: DeepSeek-R1 distilled models.
• Coding: coder versions of Qwen or other code-focused models.
• Very old or low-RAM laptops: the smallest sizes (around 1–2b parameters).