GitHub previews image generation and editing inside Copilot
GitHub announced conversational image generation and editing in GitHub Copilot on October 28. The capability is in public preview, rather than generally available, and is offered in VS Code, the GitHub Copilot app and Copilot CLI.
When a user asks for an image, the selected reasoning model calls a dedicated image-generation tool powered by GPT Image 2.5 Flare. GitHub says the result is previewed inline and saved to the project workspace.
The feature supports describing a new image or editing one through natural-language requests. Its purpose is to reduce the need to move between a coding session and a separate image tool. That product description does not establish that generated assets are accurate or suitable for every production use.
Organization and enterprise owners can disable image generation through Copilot model policies by turning off the image-generation model. Availability to a subscriber therefore does not mean an organization must permit the feature.
A visible Made with AI watermark is available as an opt-in setting for generated and edited images. Watermarking is off by default, so users should not assume every Copilot-generated image will visibly identify itself as AI-made.
The public preview covers Copilot Pro, Pro+, Business and Enterprise. Image generation is billed separately from the chat model: image tokens count toward the plan's AI credits. The announcement should not be read as a promise of unlimited image creation at no additional usage cost.