THE SIGNAL IN ONE SENTENCE
GitHub retired several AI models from Copilot on September 1, so developers and organizations may need to select replacements and retest the workflows built around them.
01
WHAT ACTUALLY CHANGED
GitHub scheduled the September 1 deprecation of Gemini 3.1 Pro, several Claude Opus and Sonnet 4.x models, Raptor Mini, and related entries in Copilot. Its notice recommends newer replacement models and describes a limited exception for some annual individual-plan users.
A model deprecation means the named model stops being available through that product, even if the surrounding chat, editor, or agent interface remains. The workflow can look unchanged while its speed, coding style, context handling, cost, or willingness to use tools shifts underneath it.
Organizations with model policies have another layer to check. A recommended replacement may exist in Copilot but still be disabled by an administrator, unavailable on a plan, or inappropriate for a sensitive repository. The migration is both a technical test and a governance task.
02
WHY THIS MATTERS
Teams increasingly save prompts, instructions, and agent routines as if they were ordinary software. But the model is a dependency with a shorter and less predictable support life than many libraries. A quiet provider change can alter a workflow without changing one line in the repository.
The sensible response is not to freeze every model forever. It is to make important workflows portable and measurable. Keep representative tasks, expected outputs, latency limits, and review criteria so a replacement can be tested before the old model vanishes.
03
WHERE IT COULD HELP
- Inventory saved prompts and agents pinned to retiring models
- Run representative coding tasks against the suggested replacements
- Check organization model policies and plan availability before migration day
- Record the model used in important automated work so changes are traceable
KEEP A HAND ON THE WHEEL
The retirement list and replacement guidance are specific to GitHub Copilot. Availability can vary by plan and organizational policy, and a replacement model should not be assumed to behave identically.
04
TERMS WORTH KEEPING
OPEN GLOSSARY CARD
Agent session
One continuing record of an agent task, including its context, actions, and results.
OPEN GLOSSARY CARD
Model deprecation
The scheduled retirement of a model from a product or service.
OPEN GLOSSARY CARD
Model policy
An organization rule that determines which AI models people and systems may use.
SOURCES AND VERIFICATION STATUS
This article was written from the materials below. Product claims and dates were checked against those sources on September 1, 2026.
PUBLICATION RECEIPT: Revision 1. Approved by Zak and published September 1, 2026.
