GitHub said on May 8, 2026 that Grok Code Fast 1 will be deprecated across all GitHub Copilot experiences on May 15, 2026. The retirement applies to Copilot Chat, inline edits, ask and agent modes, and code completions.
GitHub lists GPT-5 mini and Claude Haiku 4.5 as suggested alternatives. The company says the accelerated timing is tied to a model-provider deprecation from xAI, not only to an internal Copilot packaging decision.
Why this matters
Copilot model churn is now an operational issue for teams, not a preference-menu detail. A model can disappear from chat, completions, agent mode, and code review within a week, so enterprise administrators need a model-policy owner and a fallback plan.
It also weakens the idea that third-party model access inside Copilot is stable by default. The value of Copilot’s picker is breadth, but the cost of that breadth is more dependency on external model-provider roadmaps.
Buyer take
If your team uses Grok Code Fast 1 in Copilot, move routine coding workflows to GPT-5 mini, Claude Haiku 4.5, Sonnet 4.6, GPT-5.5, or another supported model before May 15.
For Copilot Enterprise, check model policies rather than assuming developers can switch on their own. If the suggested alternative is disabled at the organization level, users may hit broken workflows or confusing model-selection behavior.
What is still unclear
GitHub did not publish usage share for Grok Code Fast 1 inside Copilot or say whether xAI will replace it with a newer Grok coding model in the Copilot picker.
Sources
Primary and corroborating references used for this news item.