OpenAI is upgrading its widely used ChatGPT “deep research” tool, introducing a new connector that allows the AI to analyze and interact directly with GitHub codebases. This marks the company’s first external connection for the deep research mode, which compiles detailed findings on topics by searching across the web and various resources.
Rolling out initially to ChatGPT Plus, Pro, and Team subscribers, the new GitHub integration—in beta form—lets developers query specific aspects of their projects, delve into engineering documentation, and examine detailed coding structures. OpenAI plans to introduce this feature shortly thereafter to Enterprise and education-sector users.
With this connector, ChatGPT deep research will help users break down complex product specifications into manageable technical tasks and reveal dependencies within the codebase. Developers will also be able to quickly summarize coding patterns, understand implementation strategies for new APIs, and receive explicit code examples tailored from real GitHub repositories.
OpenAI describes the addition as a productivity enhancement, aimed at saving valuable time rather than substituting professional developer expertise. “We regularly hear how useful ChatGPT’s deep research engine is, and our users frequently request broader integration with their internal tools,” explained Nate Gonzalez, OpenAI’s head of business products. “Today marks an exciting step forward in interactive research capabilities.”
While recognizing the inherent limitations of any AI—including potential inaccuracies—the company emphasized that its tool will always respect organizational access privileges. Users will only see content from repositories and resources already available to them on GitHub and those explicitly shared with ChatGPT.
This update is part of OpenAI’s ongoing investment in AI-driven developer tools. Recently, the company unveiled Codex CLI, an open-source terminal coding assistant, and rolled out an upgraded ChatGPT desktop application capable of analyzing code within multiple popular software development environments. In further evidence of OpenAI’s strategic focus on coding-related technologies, reports emerged that the firm has agreed in principle to acquire AI coding startup Windsurf for around $3 billion.
Also announced on Thursday were additional fine-tuning capabilities for OpenAI’s popular language models. Developers can now customize the performance of OpenAI’s o4-mini “reasoning” model through reinforced fine-tuning and similarly fine-tune GPT-4.1 nano models, the latter being available broadly to paying developers.
These model customization tools were recently placed behind a verification requirement by OpenAI, part of an effort to ensure responsible usage, limit potential misuse, and help maintain platform integrity by requiring organizations to submit proof of identity for access.
