GitHub Copilot Gets GPT-5.3-Codex With 25% Faster AI Coding Performance

GitHub is starting the rollout of GPT-5.3-Codex, OpenAI’s newest coding-focused AI model, bringing improved performance and reasoning to GitHub Copilot across multiple platforms.

As reported by Neowin, the new model arrives shortly after the launch of Claude Opus 4.6 and OpenAI’s own GPT-5.3-Codex announcement, signaling an increasingly competitive push in advanced AI-assisted software development.

GPT-5.3-Codex comes to GitHub Copilot

According to GitHub, GPT-5.3-Codex achieved higher internal benchmark scores, with stronger reasoning and execution in complex, long-running, tool-driven workflows. The company says the model delivers around 25% faster performance on agentic coding tasks compared to GPT-5.2-Codex.

The model targets developers working on large codebases, automated refactoring, and complex software architecture, where multi-step reasoning and lower latency matter most.

Availability and supported platforms

Access to GPT-5.3-Codex requires a Copilot Pro, Pro+, Business, or Enterprise subscription. Users can select the model from the Copilot model picker in Visual Studio Code, where it works across chat, ask, edit, and agent modes.

GitHub also confirmed availability on:

• github.com
• GitHub Mobile for iOS and Android
• GitHub CLI
• GitHub Copilot Coding Agent

The rollout is gradual, so not all users will see the new model immediately. For Copilot Business and Enterprise plans, administrators must explicitly enable the GPT-5.3-Codex policy in Copilot settings.

GitHub highlights that GPT-5.3-Codex is optimized for agentic workflows, reducing latency in large-scale, tool-driven AI code generation. This makes it better suited for scenarios where Copilot needs to plan, execute, and iterate across multiple tools and files.

Developers can submit feedback through the GitHub Community, while official documentation and technical guides are expected to receive updates following the broader rollout.

Separately, OpenAI has also begun testing ads in ChatGPT, marking another notable shift in how its AI products may evolve.