YouTip

# Cursor Model Settings The power of Cursor lies in its full utilization of large language model capabilities. Cursor has many built-in large models, and of course, we can customize model settings, which is very simple. ### 1. Open Settings Click the gear icon in the upper right corner to open the settings interface. Select Models, and you can see that some large models have been set by default. Select the model you want, such as GPT-3.5 or GPT-4. We can add a model using the Add Model button at the bottom of the list. !(#) After clicking the button, an input box will appear. Enter the model name, such as deepseek-v3 and deepseek-r1. **Note:** The model name must be entered correctly without any typos. !(#) After adding the model, we need to set the API key and request URL. Enter them in the input boxes below: !(#) The API key and request URL need to be obtained from the official website of the large model. Of course, if a third-party platform provides them, you can also use those. For example, for DeepSeek, you can apply at the official website [https://platform.deepseek.com/usage](https://platform.deepseek.com/usage): !(#) !(#) The request URL is explained in the official API documentation. For DeepSeek, it is: https://api.deepseek.com/chat/completions. After completing the settings, we can click the Verify button to check if the settings are successful. If there are no error messages, it means success. The model we added will also appear in the model list. We can uncheck the other options and check only these two to start using it. !(#) Next, let's try the chat feature. Click the small icon in the upper right corner to open the chat window, then select the deepseek-r1 model, and you can start: !(#) AI is thinking... [](#)(#) [Cursor code/cursor Command](#)[](#)