Changing the AI model for Copilot code completion

Learn how to change the default LLM for Copilot Chat to a different model.

In diesem Artikel

Overview

By default, Copilot code completion uses the GPT 3.5 Turbo large language model (LLM). As an alternative, if you use VS Code, you can choose to use a model based on GPT 4o-mini. This model has been trained on a wide range of high quality public GitHub repositories, providing coverage of over 30 programming languages. Its knowledge base is more current than the default model and you may find that it generates completion suggestions more quickly.

Note

  • Multiple model support for Copilot code completion is in public preview and is subject to change.
  • Currently the ability to switch AI model is available in the latest releases of VS Code and JetBrains IDEs, with the latest version of the GitHub Copilot extension.

Effects of switching the AI model

Changing the model that's used for Copilot code completion does not affect the model that's used by Copilot Chat. See Ändern des KI-Modells für Copilot Chat.

There are no changes to the data collection and usage policy if you change the AI model.

If you are on a Copilot Free subscription, all completions count against your completions quota regardless of the model used. See Abonnementpläne für GitHub Copilot.

The setting to enable or disable suggestions that match public code are applied irrespective of which model you choose. See Suchen von öffentlichem Code, der mit GitHub Copilot-Vorschlägen übereinstimmt.

Enabling the model switcher

If you have a Copilot Free or Copilot Pro subscription, the model switcher for Copilot code completion is automatically enabled.

If you have a Copilot Business subscription, then the organization from which you receive your subscription must enable the Editor preview features setting. See Verwalten von Richtlinien für Copilot in Ihrer Organisation.

Changing the AI model for code completion

The following instructions are for VS Code. If you are using a JetBrains IDE, click the appropriate tab at the start of this article.

  1. Open the command palette by pressing Ctrl+Shift+P (Windows/Linux) / Command+Shift+P (Mac).
  2. Type change completions model and select the "GitHub Copilot: Change Completions Model" command.
  3. In the dropdown menu, select the model you want to use.

Alternatively, if Command Center is enabled, you can click the down arrow beside the icon at the top of the VS Code window, then click Configure Code Completions in the dropdown menu. Then choose Change Completions Model in the dropdown menu and select the model you want to use.

Checking which model is being used

  1. Open the Settings editor by pressing Ctrl+, (Linux/Windows) / Command+, (Mac).

  2. Type copilot completion and look for the "GitHub > Copilot: Selected Completion Model" section.

    The field in this section displays the currently selected model. If the field is empty, the default model is being used.

Changing the AI model for code completion

The following instructions are for JetBrains IDEs. If you are using VS Code, click the appropriate tab at the start of this article.

  1. Click the icon in the status bar.
  2. In the popup menu, click Edit Model for Completion.
  3. In the settings dialog box for "Languages & Frameworks > GitHub Copilot," click the dropdown menu for Model for completions and select the model you want to use.
  4. Click OK.