GitHub will retire six AI models from its Copilot services on December 14, 2026, requiring some developers and enterprise administrators to update their preferred models and integrations.
The company announced the changes on November 13.
The retirements affect supported GitHub Copilot experiences, including chat, inline edits, agent modes and code completions.
Six models scheduled for retirement
| Retiring model | Suggested replacement |
|---|---|
| Claude Fable 5 | Claude Fable 5.2 |
| Claude Haiku 4.5 | Claude Haiku 5.5 |
| Gemini 3.8 Flash | Gemini 4 Hydrogen |
| Gemini 4 Argon | Gemini 4 Carbon |
| Grok 4.6 | Grok 4.7 |
| Kimi K3 | Replacement not announced |
All six models share the same scheduled retirement date.
GitHub has not identified a replacement for Kimi K3 in the November 13 announcement.
Annual subscribers receive an exception
Claude Haiku 4.5 will remain available to individual GitHub Copilot subscribers on annual plans.
The exception does not apply generally to every subscriber or organisation.
Developers relying on that model should check their subscription type before assuming it will remain accessible.
Enterprise administrators may need to act
GitHub says suggested replacement models are automatically enabled for Copilot Business and Enterprise customers using the default model enablement policy.
However, administrators who disabled the global default or explicitly blocked a model may need to change their policies.
Once enabled, replacement models appear in the model selector of supported Copilot experiences.
GitHub says no action is required to remove models after their retirement date.
Migration affects model-dependent workflows
Developers who explicitly select models in their workflows or integrations may need to update those selections before December 14.
The announcement does not report performance comparisons between every retiring model and its proposed replacement.
It also does not state whether usage costs or model-specific limits will remain unchanged.
Organisations should therefore verify model availability and policy settings before migrating production workflows.
GitHub Changelog, 13 November 2026
















