How is DeepSeek-R1 for Coding? Try it right now!

There's hype, there's AI hype, and then there's DeepSeek-R1 hype:

Sure, the groundbreaking open-source large language model's chat app was the most-downloaded on Apple's App Store last week, but how is R1 for coding?

See for yourself!

To answer that, you could read posts and watch videos of others' impressions…or you could save time by trying it right now! Here are the 3 quick steps it takes to do that in Zed, the next-generation open-source code editor with out-the-box support for R1.

1. Download Zed
2. Add a DeepSeek API Key (via assistant: open configuration in the Command Palette)
3. Select DeepSeek Chat from the Model dropdown.

(You can also try DeepSeek Reasoner. It's slower, but can give better responses.)

That's it! Now you can take DeepSeek-R1 for a spin on your own code base.

Check out Zed's Assistant documentation to learn what you can do with it!

Running DeepSeek-R1 Locally (if you prefer)

You can also run DeepSeek-R1 on your own machine and then use it in Zed just like any other model. Here are the steps to do that:

1. Run one of the DeepSeek-R1 models on Ollama locally.
2. Configure Ollama in Zed
3. Select the new model from the Model dropdown, just like in the instructions above.

You can do this for as many different Ollama models as you like, not just DeepSeek ones.

Thank you!

Zed is open-source, and DeepSeek model support was a contribution from a community member. Thanks so much to @Cupnfish for opening a PR the same week that R1 was announced. That let it get into people's hands as fast as possible!

If you're new to Zed, it's a next-generation open-source code editor and supports many other models that you can easily try out and compare.

Now that you have it installed, check out the Getting Started tutorial!