GitHub Copilot Retires MAI-Code-1-Flash Model
GitHub CopilotGitHub Copilot deprecated MAI-Code-1-Flash across every Copilot surface, including Copilot Chat, inline edits, ask and agent modes, and code completions, effective September 10, 2026. Developers are directed to MAI-Code-1.1-Flash, Microsoft's follow-up small-tier coding model, as the suggested replacement. Copilot Enterprise administrators may need to explicitly enable the replacement model through their organization's model policy before their teams can select it, since no automatic migration path is provided.
Key Takeaways
- MAI-Code-1-Flash is gone as of September 10, 2026, removed from Copilot Chat, inline edits, ask mode, agent mode, and code completions simultaneously across the product.
- MAI-Code-1.1-Flash is the designated successor, already rolling out since August 2026 with native vision support and stronger instruction-following than the retired model.
- No migration is automatic for pinned configurations — any workflow or integration that explicitly names MAI-Code-1-Flash will lose access rather than being redirected.
- Enterprise and Business admins face an extra step, since the replacement model isn't necessarily enabled by default and may require a model-policy change before teams can select it.
- On Hacker News, the original MAI-Code-1-Flash launch drew 540 points and 255 comments, with commenters flagging it as less competitive on price-to-performance than open-weight alternatives like DeepSeek V4 Flash and Qwen 3.6.
Sources & Mentions
3 external resources covering this update
Model Retirement
On September 10, 2026, GitHub Copilot removed MAI-Code-1-Flash from every Copilot experience it powered, spanning Copilot Chat, inline edits, ask mode, agent mode, and code completions. The model, Microsoft's first purpose-built small-tier coding model for Copilot, is no longer selectable anywhere in the product as of the retirement date.
GitHub Copilot is pointing developers toward MAI-Code-1.1-Flash as the direct successor. That model, which GitHub had already been rolling out since August 2026, built on the original with native vision support for image understanding and improvements to coding quality, instruction following, and tool use, while shipping at a substantially lower list price than its predecessor.
Why It Matters
Model retirements like this one affect anyone who pinned MAI-Code-1-Flash specifically in a workflow, integration, or personal model preference rather than relying on Copilot's automatic model selection. Once the cutoff hit, those configurations silently lost access to the model rather than failing outright, so any automation or CI step that explicitly names MAI-Code-1-Flash needs to be updated to reference the newer model instead.
What Developers Should Do
GitHub Copilot says no action is required to remove the deprecated model itself, since it simply disappears from the interface. However, developers who want to keep using a Microsoft small-tier model for fast, high-volume, iterative coding tasks should switch their workflows and integrations to MAI-Code-1.1-Flash.
For Copilot Business and Enterprise customers, the replacement model is not necessarily available by default. Administrators may need to enable MAI-Code-1.1-Flash through their organization's Copilot model policy before it appears in the Copilot Chat model selector in VS Code, on github.com, and across other Copilot surfaces. Administrators can confirm availability by checking their individual Copilot settings and verifying the policy is enabled for that specific model. GitHub Enterprise customers with questions were directed to reach out to their account manager for assistance.