ChatGPT Deep Research Gets GPT-5.2 Update With Smarter Controls

If you rely on ChatGPT for serious research work, there’s an upgrade for you that is rolling out right now. OpenAI has announced that Deep Research inside ChatGPT is now powered by GPT-5.2, which brings workflow controls, richer outputs, and more transparency to how results are built.

OpenAI upgrades Deep Research in ChatGPT with GPT-5.2, adding advanced controls and structured reporting

Deep Research, first introduced last year, works as an agentic research layer inside ChatGPT. Rather than just answering prompts, it performs multi-step internet research for complex queries, stitching together findings, comparisons, and structured analysis in one place. Now, with GPT-5.2 under the hood, the experience is becoming far more customizable.

Users can connect supported apps directly within ChatGPT and target specific websites during research. There’s also the option to restrict results to selected domains or defined time ranges, apply filters, and set comparison parameters. Real-time progress tracking is built in, and ongoing sessions can be interrupted to insert follow-up instructions or new sources without restarting the entire workflow.

On the output side, the update introduces structured reporting tools. Research results can now be generated as organized reports or tables, complete with visual charts and data summaries. A full-screen report mode makes longer outputs easier to review, especially for analysis-heavy tasks.

OpenAI is also emphasizing transparency, as reports may include summarized key findings, visual comparisons, methodology explanations detailing how conclusions were compiled, contextual commentary, and source citations for verification. Everything remains embedded within the chat interface, meaning users can review, refine, and export without switching tools.