OpenAI has announced that Deep research in ChatGPT is now powered by GPT-5.2. Introduced last year, Deep research is an agentic capability that performs multi-step internet research inside ChatGPT for complex tasks.
The update expands how users control research workflows and how results are generated, visualized, and reviewed within the chat interface.
Research controls and output capabilities
With GPT-5.2 integration, Deep research allows users to customize both the research process and final outputs. Users can:
- Connect supported apps within ChatGPT and search specific websites
- Restrict research to selected domains or defined time ranges
- Apply filters and comparison parameters
- Track research progress in real time
- Interrupt an active session to add follow-ups or new sources
- Generate structured outputs such as reports or tables
- Include charts and visual data summaries
- View results in fullscreen report mode
These tools let users adjust how information is gathered and presented without leaving ChatGPT.
Structured reporting and transparency
Deep research produces integrated reports that combine analysis and documentation. Reports may include:
- Key findings and summarized insights
- Visual data comparisons and trend charts
- Methodology explanations outlining how results were compiled
- Commentary sections for context
- Source citations for verification
Reports remain embedded in the chat workflow, enabling review and export within ChatGPT.
Availability
The GPT-5.2 powered Deep research update is rolling out starting today.