OpenAI GPT-5.3-Codex Brings Next-Gen AI Coding to GitHub Copilot

OpenAI GPT-5.3-Codex Brings Next-Gen AI Coding to GitHub Copilot

OpenAI has released GPT-5.3-Codex, its latest agentic coding model, now generally available in GitHub Copilot. The new model delivers significant improvements in coding benchmarks and complex workflow execution.

Performance Improvements

According to GitHub's announcement, GPT-5.3-Codex reaches new high scores on benchmarks used to evaluate coding, agentic, and real-world capabilities. The model shows improved reasoning and execution in complex, tool-driven, long-running workflows, delivering up to 25% faster performance than GPT-5.2-Codex on agentic coding tasks.

Availability

GPT-5.3-Codex is available to Copilot Pro, Pro+, Business, and Enterprise users. Developers can select the model in the model picker across multiple platforms:

• •Visual Studio Code in all modes: chat, ask, edit, and agent
• •GitHub Mobile for iOS and Android
• •GitHub Copilot CLI
• •GitHub Copilot Coding Agent

The rollout is gradual, so users who don't see it yet should check back soon.

Enterprise Configuration

Copilot Enterprise and Copilot Business plan administrators must enable the GPT-5.3-Codex policy in Copilot settings to provide access to their organizations.

Context

GPT-5.3-Codex follows the release of GPT-5.3-Codex-Spark in February 2026. OpenAI has been rapidly iterating on its coding models, with GPT-5.1 launching in November 2025 and GPT-5.2 in December 2025.

This release comes amid intensifying competition in the AI coding space, with Anthropic's Claude and Google's Gemini also advancing their coding capabilities.

---

Sources:

• •GitHub Changelog: GPT-5.3-Codex now generally available
• •Wikipedia: GPT-5