GitHub Copilot is evolving rapidly, with new features and models released regularly. If you're an enterprise administrator, staying informed helps you make confident decisions about which capabilities to enable, when to adopt them, and how to manage risk across your organizations.
This article provides guidance on navigating the Copilot landscape, understanding different types of features and models, and preparing for upcoming changes.
Keeping up to date with Copilot changes
You can use GitHub Docs and the GitHub Blog to keep track of new releases.
Learning about new Copilot features
To learn about new Copilot features, we recommend monitoring these two locations:
- Features overview: For a complete list of Copilot capabilities, see fonctionnalités de GitHub Copilot.
- Changelog: Follow the Copilot changelog for announcements about new and updated features.
Copilot features generally fall into three categories:
| Type | Description | Examples | Considerations |
|---|---|---|---|
| Assistive | Respond to prompts and provide suggestions, but require human review before changes are made. | Inline suggestions, Discussion avec Copilot, résumés des demandes de tirage Copilot. | All changes require user approval. |
| Agentic | Autonomously research, plan, and make changes on behalf of users. | Agent cloud Copilot, third-party agents. | Can work autonomously, but with built-in protections. |
| Third-party | External coding agents that work alongside Agent cloud Copilot to complete tasks asynchronously. | Anthropic Claude, OpenAI Codex. | Same as agentic. Review provider documentation and enable via policies. See À propos des agents de codage tiers. |
Each feature has its own enablement requirements and policy settings. When a new feature is released:
- Review the feature documentation to understand its capabilities.
- Check the policy settings available at the enterprise and organization level. See stratégies de GitHub Copilot pour les entreprises et les organisations.
- Consider running a pilot with a subset of users before broader rollout.
Learning about new Copilot models
GitHub Copilot uses multiple AI models from different providers to power its features. Different models have different strengths: some prioritize speed and cost-efficiency, while others are optimized for accuracy, reasoning, or working with multimodal inputs (like images and code together).
Users will have access to different models depending on the feature they're using and your enterprise's Copilot plan and policies.
You can find information about the models available and upcoming models in the following locations:
- Supported models: For a complete list of available models and their capabilities, see Modèles IA pris en charge dans GitHub Copilot.
- Model comparison: To compare model capabilities side by side, see Comparaison des modèles IA.
- Changelog: Model updates are announced in the Copilot changelog.
Default enablement
On Copilot Business and Copilot Entreprise plans, a new policy will control whether unconfigured generally available (GA) models default to enabled or disabled. For more information, see About default availability of Copilot models.
Special categories
GitHub categorizes certain types of model, allowing you to plan for model transitions and set user expectations.
| Model type | Description | Why it matters |
|---|---|---|
| Utility model | A small set of models that power background GitHub features across surfaces. | Your enterprise cannot disable these models. Monitor these to ensure they are compliant. |
| Base model | The default model when no other models are enabled. | Automatically enabled within 60 days of designation. |
| LTS model | A model supported for one year from designation. | Allows enterprises to build workflows around a stable model. |
For more information, see Modèles utilitaires and Modèles de prise en charge de base et à long terme (LTS).
Considering different release stages
Copilot features and models progress through different release stages. Understanding release stages helps you decide when to enable features for your organization. Each stage has different expectations for stability and support.
| Stage | Availability | Stability |
|---|---|---|
| Préversion privée | Limited set of customers by invitation. | May change significantly based on feedback. |
| Version d'essai publique | All eligible customers who opt in. | Functional but may have limitations or change. |
| Generally available (GA) | All customers. | Fully supported and recommended for production. |
Conseil
Many preview features are covered by the GitHub Data Protection Agreement, meaning you can test them without breaching compliance requirements. See GitHub DPA-Covered Previews for a list.
Preview features are controlled by policy settings at the organization and enterprise level. The Opt in to preview features policy allows administrators to enable or disable preview features on the GitHub website specifically. See Gestion des stratégies et fonctionnalités pour GitHub Copilot dans votre entreprise and Gestion des stratégies et des fonctionnalités pour GitHub Copilot dans votre organisation.
Reviewing information about Responsible AI
GitHub is committed to responsible AI development and provides transparency documentation for each Copilot feature.
Each Copilot feature has a dedicated responsible use article describing its purpose, capabilities, and limitations. See Utilisation responsable des fonctionnalités de GitHub Copilot.