BotBeat
...
← Back

> ▌

MicrosoftMicrosoft
UPDATEMicrosoft2026-05-02

GitHub Copilot to Deprecate GPT-5.2 and GPT-5.2-Codex Models on June 1st

Key Takeaways

  • ▸GPT-5.2 and GPT-5.2-Codex models will be deprecated across most GitHub Copilot experiences on June 1st, 2026, except for GPT-5.2-Codex in Code Review
  • ▸Users must update their workflows and integrations to use supported models before the deprecation date
  • ▸Copilot Enterprise administrators need to enable and configure alternative models through their policy settings
Source:
Hacker Newshttps://github.blog/changelog/2026-05-01-upcoming-deprecation-of-gpt-5-2-and-gpt-5-2-codex/↗

Summary

GitHub announced that it will deprecate the GPT-5.2 and GPT-5.2-Codex models across most GitHub Copilot experiences (including Copilot Chat, inline edits, ask and agent modes, and code completions) on June 1st, 2026. Notably, GPT-5.2-Codex will continue to be available in Copilot Code Review. This deprecation requires developers to migrate their workflows and integrations to supported alternative models before the deadline.

Copilot Enterprise administrators will need to enable access to alternative models through their model policies in Copilot settings. Administrators can verify model availability by checking individual Copilot settings and confirming that policies are enabled for specific models. Once enabled, users will see alternative models in the Copilot Chat model selector in VS Code and on github.com.

No action is required to remove the deprecated models after the deadline passes. GitHub Enterprise customers with questions about the transition are encouraged to contact their account managers for further assistance.

Large Language Models (LLMs)AI AgentsMarket Trends

More from Microsoft

MicrosoftMicrosoft
INDUSTRY REPORT

Microsoft's AI Ambitions Put OpenAI and Anthropic on Notice

2026-07-30
MicrosoftMicrosoft
RESEARCH

Russian Espionage Group TA488 Expands Half-Click Email Attack to Microsoft Outlook

2026-07-30
MicrosoftMicrosoft
RESEARCH

Security Researcher Reveals Self-Propagating 'AI Worm' Attack in Microsoft Copilot for Word

2026-07-30

Comments

Suggested

LoreLore
PRODUCT LAUNCH

Velocity: Proof of Linear-Scaling Long Context Without Model Retraining

2026-08-02
NixNix
RESEARCH

High-Bandwidth Flash Could Scale GPU Memory to Multiple Terabytes

2026-08-02
AI Industry (General Analysis)AI Industry (General Analysis)
INDUSTRY REPORT

Damned if You Do, Damned if You Don't: Professional Negligence Liability in the Age of AI

2026-08-02
← Back to news
© 2026 BotBeat
AboutPrivacy PolicyTerms of ServiceContact Us