GitHub Copilot's Auto Model Selection Now Generally Available in Visual Studio Code

GitHub Copilot's auto model selection is now generally available within Visual Studio Code for all Copilot plans. This feature allows Copilot to intelligently choose the most suitable AI model based on real-time availability, optimizing performance and mitigating rate limits.

How Auto Model Selection Works

The auto model selection system is dynamic, routing requests to available models such as GPT-5.1-Codex-Max, GPT-5 mini, GPT-4.1, Sonnet 4.5, and Haiku 4.5. The specific models utilized are determined by your Copilot plan and established policies, and this selection may evolve over time.

Key benefits include:

• Transparency: Users can easily identify which model was used for a response by hovering over it.
• User Control: The flexibility to switch between the automatic selection and a specific model is maintained at all times.
• Policy Compliance: Auto model selection fully respects all user and administrator model settings and policies.

Premium Request Usage

Premium requests, when using auto model selection, are billed based on the specific model chosen by the system. Currently, this is limited to models with 0x to 1x multipliers. All paid subscribers benefit from a 10% discount on the model multiplier when using auto. For example, if auto selects a model with a 1x multiplier, your usage will be recorded as 0.9 premium requests instead of 1.

The Road Ahead

While auto currently routes to readily available, high-quality models, its capabilities are set to expand. In the near future, auto will become even more intelligent, gaining enhanced features that will enable Copilot to select the most appropriate model for your task, precisely matching the model to the complexity level of your request.

For questions or to provide feedback, please join the community discussion.