{
  "altLabels": [
    "Agentic Deep Research"
  ],
  "definition": "An agentic information-seeking pattern in which a system performs multi-step research by planning searches, browsing or retrieving sources, reasoning over evidence, and synthesizing a cited answer or report.",
  "derivedRelations": [
    {
      "target": "browsecomp",
      "type": "evaluatedBy"
    },
    {
      "target": "browsecomp-plus",
      "type": "evaluatedBy"
    },
    {
      "target": "browsecomp",
      "type": "related"
    },
    {
      "target": "browsecomp-plus",
      "type": "related"
    }
  ],
  "distinctions": [
    {
      "reason": "Agentic RAG is a generation pattern with agentic retrieval or tool-use control; deep research emphasizes broader multi-step research and cited synthesis.",
      "statement": "Deep Research is not Agentic RAG.",
      "target": "agentic-rag",
      "targetLabel": "Agentic RAG",
      "targetUri": "https://id.searchplex.net/agentic-rag/"
    },
    {
      "reason": "Deep research is a broader agentic research and synthesis pattern; agentic retrieval is the retrieval loop that may support it.",
      "statement": "Deep Research is not Agentic Retrieval.",
      "target": "agentic-retrieval",
      "targetLabel": "Agentic Retrieval",
      "targetUri": "https://id.searchplex.net/agentic-retrieval/"
    },
    {
      "reason": "Question answering returns an answer to a question; deep research often produces a synthesized, cited research result after multiple search and reasoning steps.",
      "statement": "Deep Research is not Question Answering.",
      "target": "question-answering",
      "targetLabel": "Question Answering",
      "targetUri": "https://id.searchplex.net/question-answering/"
    },
    {
      "reason": "RAG is a retrieval-augmented generation pattern; deep research emphasizes multi-step search, browsing, evidence gathering, and cited synthesis.",
      "statement": "Deep Research is not Retrieval-Augmented Generation.",
      "target": "rag",
      "targetLabel": "Retrieval-Augmented Generation",
      "targetUri": "https://id.searchplex.net/rag/"
    }
  ],
  "id": "deep-research",
  "kind": "GeneralConcept",
  "license": "https://creativecommons.org/publicdomain/zero/1.0/",
  "links": [
    {
      "label": "Introducing deep research",
      "type": "officialHomepage",
      "url": "https://openai.com/index/introducing-deep-research/"
    },
    {
      "label": "From Web Search towards Agentic Deep Research",
      "type": "canonicalSurvey",
      "url": "https://arxiv.org/abs/2506.18959"
    },
    {
      "label": "DeepResearch Bench",
      "type": "definingReference",
      "url": "https://arxiv.org/abs/2506.11763"
    }
  ],
  "modifiedAt": "2026-09-14T06:43:46Z",
  "prefLabel": "Deep Research",
  "publisher": {
    "name": "Searchplex",
    "url": "https://searchplex.net/"
  },
  "relations": [
    {
      "note": "Agentic RAG is a generation pattern with agentic retrieval or tool-use control; deep research emphasizes broader multi-step research and cited synthesis.",
      "target": "agentic-rag",
      "type": "notEquivalentTo"
    },
    {
      "note": "Deep research is a broader agentic research and synthesis pattern; agentic retrieval is the retrieval loop that may support it.",
      "target": "agentic-retrieval",
      "type": "notEquivalentTo"
    },
    {
      "note": "Question answering returns an answer to a question; deep research often produces a synthesized, cited research result after multiple search and reasoning steps.",
      "target": "question-answering",
      "type": "notEquivalentTo"
    },
    {
      "note": "RAG is a retrieval-augmented generation pattern; deep research emphasizes multi-step search, browsing, evidence gathering, and cited synthesis.",
      "target": "rag",
      "type": "notEquivalentTo"
    },
    {
      "target": "agentic-rag",
      "type": "related"
    },
    {
      "target": "agentic-retrieval",
      "type": "related"
    },
    {
      "target": "multi-hop-retrieval",
      "type": "related"
    },
    {
      "target": "query-decomposition",
      "type": "related"
    },
    {
      "target": "question-answering",
      "type": "related"
    },
    {
      "target": "rag",
      "type": "related"
    },
    {
      "target": "retrieval-planning",
      "type": "related"
    },
    {
      "target": "web-browsing-agent",
      "type": "related"
    },
    {
      "target": "web-browsing-agent-evaluation",
      "type": "related"
    }
  ],
  "representations": {
    "html": "https://id.searchplex.net/deep-research/",
    "json": "https://id.searchplex.net/deep-research.json",
    "jsonld": "https://id.searchplex.net/deep-research.jsonld"
  },
  "scheme": "https://id.searchplex.net/scheme/",
  "scopeNote": "Deep research usually combines agentic retrieval, web browsing, source evaluation, citation, and synthesis. It is related to RAG and question answering, but it emphasizes multi-step research behavior and report-style synthesis rather than a single retrieval or generation step.",
  "status": "published",
  "terms": [
    {
      "community": "industry",
      "label": "Deep Research Agent",
      "usage": "nearSynonym"
    },
    {
      "community": "academic",
      "label": "DRA",
      "usage": "abbreviation"
    }
  ],
  "uri": "https://id.searchplex.net/deep-research/"
}