Deepseek R1 is a newly released open-source AI model that offers impressive performance in math, coding, and reasoning tasks. The R1 model, akin to GPT-1, offers enhanced generative capabilities, enabling the creation of text, answering questions, and facilitating creative and routine tasks across diverse applications such as education, business, and content creation. Deepseek R1 has proven to be as effective as OpenAI’s GPT-4o and Claude 3.5 Sonnet in various benchmarks.
One of its standout features is the ability to run locally/offline after the initial setup, unlike other models that require a constant internet connection. This makes it a great option for users who need flexibility and privacy.
Below is a step-by-step guide to running a simplified/distilled version of Deepseek R1 through your computer’s terminal. (Note: This guide is specifically for Windows 11 users.)
1) Install Ollama
Quick intro to Ollama: It’s a simple tool for running AI models locally on your machine.
Download it here: https://ollama.com/download.
2) Next, you’ll need to run the DeepSeek R1 model locally.
Ollama offers various model sizes—larger models deliver more advanced AI capabilities but require a more powerful GPU to run efficiently.
1.5B version (smallest):
ollama run deepseek-r1:1.5b
8B version:
ollama run deepseek-r1:8b
14B version:
ollama run deepseek-r1:14b
32B version:
ollama run deepseek-r1:32b
70B version (biggest/smartest):
ollama run deepseek-r1:70b
Note: Larger models like 32B and 70B require significant GPU power. Start with a smaller model and scale up based on your hardware capabilities!
Start with a smaller model to get a feel for how it works. Simply open your terminal and run:
The model will start downloading automatically.
Once downloaded, the model will run locally on your machine—it’s that simple!
Althought users may not be accustomed to interacting with a program through the terminal, this method is the simplest way to start experimenting with your own AI models at home without too much hassle.
AI for Success 