• Guide

How to download and install DeepSeek AI

There are a number of ways to utilize DeepSeek AI, and one of the most interesting is to download and install it locally.

Aidan O'Brien

January 28, 2025 6:00 AM

DeepSeek

1

DeepSeek AI is causing quite the buzz in the AI space, knocking value off companies like NVIDIA and generally getting the nascent market to rethink exactly what is required to develop and run these models. If you are interested in experimenting with DeepSeek for yourself, this is how you can do it. 

How to download and install DeepSeek AI

DeepSeek is available on both the Apple and Android stores as DeepSeek Assistant. This assistant is based on DeepSeek-V3, which, it should be noted, is not the DeepSeek R1 model that has caused such a stir. DeepSeek is also available in a browser-based model, much like ChatGPT. The reason I mention these is that it is likely you will need to use these versions if you do not have a machine that is good enough for local installation.

The good news is that local installation has been made very easy, thanks to Ollama. Head to the Ollama download page and select which operating system you use. Ollama is available on MacOS, Linux, and Windows. Run the Ollama terminal once it is installed, and type in one of the following commands, depending on which model you wish to run:

• ollama run deepseek-r1:1.5b
• ollama run deepseek-r1:7b
• ollama run deepseek-r1:8b
• ollama run deepseek-r1:14b
• ollama run deepseek-r1:32b
• ollama run deepseek-r1:70b

The end number indicates how heavy the model will be to run, and while the lighter models will not be as good as the heavier ones, they still have fantastic performance, likely more than good enough for a casual level of exploration. Needless to say, the heavier the models you wish to run, the more demanding that would be, and the more space you will need for the model installation.

I would strongly suggest you start at the 1.5b model and work your way up, regardless of your machine's specs. 

Now that we have Ollama and our model installed, we need a way to interact with it all, something that has once again been made very easy thanks to Chatbox AI. Check the pricing to ensure you are fully informed on what you might want to do, then head to the download page and grab the suitable version for your operating system.

Install Chatbox AI and run it, then click the Chatbox AI button under where it says Model Provider. Select Ollama, then in Model, choose the same deepseek-r1 model that you installed using the above command. 

There you have it; you can now run and interact with DeepSeek-R1 on your system. We'll be keeping an eye on DeepSeek to see if it continues to shine, so stick with Shacknews for more DeepSeek news and guides.

Aidan O'Brien

Contributing Guides Editor

Hailing from Ireland, Aidan has been conditioned by local weather conditions to survive hours at his PC grinding through whatever game is offering the lowest possible drop rates for loot. He thinks the easiest way to figure out what fans of games want to read is to just be a fan of games. You can normally find him logged into Warframe, Destiny, or a gacha game. You can reach out to him on X @scannerbarkly.