On this page
GitHub’s Changelog dated October 28, 2026 says GitHub Copilot can generate and edit images with natural language. The stage is public preview, not generally available. An image request causes the selected reasoning model to call a dedicated image-generation tool powered by GPT Image 2.5 Flare, preview the result inline, and save it to the project/workspace. That sits beside, not inside, the chat-model menu already covered here: Which Models GitHub Copilot Offers Right Now, and What Each One Costs.
Surfaces and controls
Image generation is available in a Copilot chat session in Visual Studio Code, the GitHub Copilot app, or Copilot CLI, for a new request or mid-coding task. Creating a new asset or editing something already in the thread uses the same image tool path. Organization and enterprise owners can turn off image generation by disabling the image-generation model through Copilot model policies. A visible “Made with AI” watermark on images Copilot generates and edits is available as an opt-in; watermarking is off by default. The Changelog does not claim Free-plan access for this preview.
Plans and billing
Image generation is in public preview for Copilot Pro, Copilot Pro+, Copilot Business, and Copilot Enterprise. Usage is billed separately from the chat model, based on image tokens that count toward the plan’s AI credits. How credits work on this site: GitHub Copilot AI Credits: GitHub Now Bills by the Token, and Your Model Choice Sets the Price. Image spend and chat routing are separate meters in the Changelog.
A different product’s image-generation news (ChatGPT ads test, not Copilot): OpenAI to Test Visual Ads Inside ChatGPT Image Generation.
This is a public preview announcement, not a generally available claim. Surfaces, policies, and billing can differ by account. Details follow the GitHub Changelog. For reference only.