OpenAI Deep Research: the agent that conducts deep research for tens of minutes

Doing thorough research on a complex topic takes hours: opening dozens of tabs, reading articles, cross-referencing sources, synthesizing contradictory information. Deep Research automates exactly this process.

Based on the o3 model, the agent receives a complex question, autonomously plans search queries, browses the web for 10-30 minutes, reads hundreds of pages and sources, and produces a detailed report with verifiable citations and links to original sources.

It's not just a summary: it's a multi-source analysis with checkable references, similar to what a human analyst would produce after half a day of work. On GPQA (expert research benchmark), Deep Research scores comparably to PhD-level specialists.