Why it matters
This deprecation impacts developers relying on specific AI models within GitHub Copilot for code generation and assistance. Understanding the transition to newer models is crucial for maintaining seamless development workflows and leveraging the latest AI capabilities.

What changed

As of June 5, 2026, GitHub Copilot has deprecated the GPT-5.2 and GPT-5.2-Codex models across most of its features, including Copilot Chat, inline edits, ask, agent modes, and general code completions. The GPT-5.2 model remains accessible specifically for Copilot code review functionalities. Developers are directed to update their integrations and workflows to utilize the suggested alternative models: GPT-5.5 as a replacement for GPT-5.2, and GPT-5.3-Codex for GPT-5.2-Codex.

For GitHub Enterprise customers, administrators have the ability to enable access to these alternative models through their model policies within Copilot settings. Verification of model availability can be done via individual Copilot settings, ensuring the policy is active for the desired model. Once enabled, the new models will appear in the Copilot Chat model selector in VS Code and on github.com. No explicit action is needed to remove the deprecated models from existing configurations.

Why it matters for builders

Developers using GitHub Copilot need to be aware of these model changes to ensure their tools continue to function as expected. Migrating to the recommended GPT-5.5 and GPT-5.3-Codex models will be necessary for accessing the full suite of Copilot's AI-powered coding assistance features. This transition requires attention to workflow adjustments and potential policy updates for enterprise users.

Practical impact

Builders should proactively update any custom scripts, integrations, or configurations that explicitly reference GPT-5.2 or GPT-5.2-Codex. For those using Copilot Enterprise, administrators should review and update model policies to enable GPT-5.5 and GPT-5.3-Codex. Checking the Copilot documentation for the latest model information and guidance on migration is recommended. Developers can verify model availability in their VS Code or github.com Copilot Chat interfaces after administrator policy updates.

Caveats and source limits

The source provides a clear deprecation date and suggested alternatives but does not offer performance benchmarks or detailed feature comparisons between the deprecated and new models. Information regarding specific API changes or migration guides beyond updating model selection is not detailed. The exact scope of availability for GPT-5.2 in code review is also not elaborated upon.

Sources

Written with AI assistance from the linked sources; every claim below was checked against them automatically. How we produce articles.

Claim check: 5/5 supported claims - 5 evidence links - 100% avg confidence
  • GPT-5.2 and GPT-5.2-Codex models have been deprecated across most GitHub Copilot experiences.supported - github.blog
  • The deprecation date for GPT-5.2 and GPT-5.2-Codex was June 5, 2026.supported - github.blog
  • GPT-5.5 is suggested as an alternative to GPT-5.2.supported - github.blog
  • GPT-5.3-Codex is suggested as an alternative to GPT-5.2-Codex.supported - github.blog
  • GPT-5.2 remains available for Copilot code review.supported - github.blog

Caveats

  • Single-source caution: verify critical details at the linked source.
Radar score 74/100 - how it was calculated
Reliability87
Freshness85
Novelty47
Technical45
Developer56
Ecosystem78
Confidence100
  • Reliability 87: Primary official source
  • Freshness 85: Fresh official source date
  • Novelty 47: Official announcement
  • Technical 45: Technical release details
  • Developer 56: Developer-facing announcement
  • Ecosystem 78: Official source
  • Confidence 100: Claims have reliable evidence
Share
XLinkedInHacker News

Related articles

Enterprise AI - Sep 15, 2026Fyxer AI Assistant Uses OpenAI Models for Email ManagementFyxer, an AI executive assistant, leverages OpenAI models, fine-tuning, and memory capabilities to manage user inboxes and draft emails. The system incorporates real user feedback to personalize email responses in the user's distinct voice.Enterprise AI - Aug 14, 2026OpenAI's Daybreak Cybersecurity Models Available on AWS BedrockOpenAI's Daybreak cybersecurity models are now accessible via Amazon Bedrock. This integration aims to enhance enterprise security workflows by providing specialized AI capabilities.Developer Tools - Sep 29, 2026Open Source AI Radar: AI Project Intelligence PlatformOpen Source AI Radar is a new intelligence platform designed to discover and analyze emerging open-source AI projects on GitHub. It utilizes a three-axis scoring system (Impact, Velocity, Health) to identify projects gaining importance, moving beyond simple star counts.Developer Tools - Sep 29, 2026AFDocs v0.22.1: Test Documentation for AI Agent CompatibilityAFDocs, a new tool from agent-ecosystem, allows developers to test their documentation against the Agent-Friendly Documentation Spec. The latest release, v0.22.1, checks for discoverability, navigation, and consumption by AI coding agents, providing a scorecard with detailed category scores and fix suggestions.Developer Tools - Aug 17, 2026Agent Deck: Terminal Session Manager for AI Coding AgentsAgent Deck is a terminal-based session manager designed for AI coding agents. It provides a unified TUI interface for interacting with various AI models like Claude, Gemini, and Codex.Developer Tools - Sep 29, 2026Seiso: Markdown Convention and Linter for AI-Generated DocsSeiso is a new Rust-based Markdown convention and linter designed to standardize project documentation written by AI. It aims to improve consistency and accuracy for both human readers and AI agents by enforcing a structured format and identifying discrepancies.