VS Code (GitHub Copilot Chat) Setup

Installation

Download and install VS Code from the VS Code website, then search for and install the GitHub Copilot Chat extension from the marketplace.

Configure a Custom Model Endpoint

1. Open the Chat panel, click the model picker button next to the input box, and select Manage Models from the menu.

2. Click Add Model, then choose Custom Endpoint.

3. Enter a group name (the group name shown in the model picker, e.g. Tokenfly).

4. Enter a display name and API Key: sk-xxx.

5. Select Chat Completions as the API type.

6. VS Code will automatically open chatLanguageModels.json — fill in the model details:

   [
     {
       "name": "Tokenfly",
       "vendor": "customendpoint",
       "apiKey": "sk-xxx",
       "apiType": "chat-completions",
       "models": [
         {
           "id": "gpt-5.4",
           "name": "GPT-5.4",
           "url": "https://tokenfly.com/v1/chat/completions",
           "toolCalling": true,
           "maxInputTokens": 900000,
           "maxOutputTokens": 100000
         }
       ]
     }
   ]

7. Save the file, then select GPT-5.4 in the Chat panel's model picker to use it.

Configuration Reference

Verification

Select GPT-5.4 in the model picker and start a conversation — a normal reply confirms the setup works. If the newly added model doesn't appear immediately in the picker, restart VS Code.