Research
Claude's Research mode runs many web searches that build on each other, plus searches of connected tools such as Gmail and Google Docs, and returns a cited report. It is on paid Claude plans (Pro, Max, Ultra, Team, Enterprise), not Free.
Research is Claude's deep research mode: you turn it on with the "+" button and "Research", and Claude runs a series of searches that build on each other across the web and your connected tools, then writes an answer with citations; it is on Pro, Max, Ultra, Team and Enterprise, not Free (as of Nov 6, 2026). Anthropic launched it on Apr 15, 2025 as an early beta for Max, Team and Enterprise in the United States, Japan and Brazil, promising answers "in minutes". The Help Center gives no fixed running time today; Anthropic's Dec 2, 2025 getting-started video puts most reports at 5 to 15 minutes and the most complex at up to 45 (Getting started with research in Claude.ai).
The closest rival is ChatGPT's deep research, which OpenAI includes on ChatGPT's Free plan while Claude's needs a paid plan; which writes the better report has no independent, dated measurement to settle it. Learn more: Deep Research on chatgpt.vi.ki
Quick facts
| Plans | Pro, Max, Ultra, Team, Enterprise (as of Nov 6, 2026) |
| Apps | Web, Claude Desktop, Claude Mobile |
| Model | Claude Opus 5.5, falling back to Claude Opus 4.8, whatever the chat uses |
| Sources | The web plus connected tools (Gmail, Google Calendar, Google Docs, other connectors) |
| Launched | Apr 15, 2025 (early beta) |
How do I use Research in Claude?
- Make sure web search is on: Research needs it.
- In a new chat, click "+" at the bottom left.
- Choose "Research". A blue indicator shows it is on.
- Ask a question that needs digging: a market overview, a comparison of options, a literature summary, "what changed in this regulation this year".
- Wait while Claude searches, reads and plans its next searches, then read the report and check its citations.
What Research does differently from web search
Ordinary web search answers from one or a few searches. Research runs several searches in a row, each informed by the last, decides which angles to explore next and works through open questions before it answers (Use research on Claude). Every claim in the result carries a citation you can open. The trade is time and usage: a Research run takes longer than a chat reply, and the Help Center warns it can use up your usage limits faster.
Which model does Research use?
Research always runs on Claude Opus 5.5 and falls back to Claude Opus 4.8, whichever model you picked for the chat (Use research on Claude). On Team and Enterprise, if an admin has turned off both models for your role, Research fails rather than using another model.
Research also thinks before it answers: Anthropic's Dec 2, 2025 video says extended thinking is switched on automatically with Research, shown as a timer while Claude works and an expandable thinking section above the reply.
Research with Gmail, Calendar and Google Docs
Research can search your own data through Connectors enabled in that chat: Gmail, Google Calendar, Google Docs and others, which it uses without asking each time (Use research on Claude). This was part of the launch: the Apr 15, 2025 post paired Research with a Google Workspace integration that searches email, documents and calendar commitments without uploads (Claude takes research to new places). On Team and Enterprise an admin enables Google Workspace for the domain first.
Two gaps the Help Center states plainly: a Research run started in the mobile app searches only the web and skips connectors without warning, and Research cannot use local MCP servers set up in the Claude Desktop App.
Frequently asked questions
Is Claude Research free? No. Claude Research is on the Pro, Max, Ultra, Team and Enterprise plans as of Nov 6, 2026; the Free plan has web search but not Research.
How long does Claude Research take? Anthropic's Dec 2, 2025 video says most Claude Research reports take 5 to 15 minutes and complex ones up to 45; the Help Center gives no fixed duration, and a run takes longer than an ordinary reply because Claude searches several times in a row.
Does Claude Research cite its sources? Yes. Every Claude Research report carries citations that link to the pages or documents each claim came from.
Why did Claude Research ignore my Gmail on my phone? A Research run started in the Claude mobile app searches only the web and skips connectors such as Gmail, with no warning; start it on the web or in Claude Desktop to include them.
Is Claude Research the same as ChatGPT deep research? They do the same job: many searches, then a cited report. ChatGPT offers deep research on its Free plan, while Claude Research needs a paid plan; see Claude vs ChatGPT.
Sources
- Getting started with research in Claude.ai, Anthropic, Dec 2, 2025 (video)
- Use research on Claude, Claude Help Center, read Nov 6, 2026
- Claude takes research to new places, claude.com, Apr 15, 2025
- Plans and pricing, claude.com, read Nov 6, 2026
- Plans and Pricing - ChatGPT, OpenAI, read Nov 6, 2026
Moments
Getting started with research in Claude.ai
Anthropic
- Research is Claude's deep research mode: you turn it on with the "+" button and "Research", and Claude runs a series…
- Ordinary web search answers from one or a few searches. Research runs several searches in a row, each informed by the…
- Research also thinks before it answers: Anthropic's Dec 2, 2025 video says extended thinking is switched on…
Photos
See also
-
Web SearchYes, Claude can search the internet. Web search is on every Claude plan, Free included, in every country Claude serves; turn it on with "+" and "Web search", and every answer that uses it cites its sources.
-
Claude Usage LimitsClaude's paid plans have a session limit that resets five hours after your first message and a weekly limit that resets at a fixed time each week; Anthropic publishes multiples of Pro's allowance, not message counts.
-
Claude Opus 5.5Claude Opus 5.5 is Anthropic's recommended starting model, released on Sep 22, 2026 at $4 input and $20 output per million tokens on the Claude API, with a 1M-token context window. It is on every paid Claude plan and is the default model in Claude Code. -
Extended and Adaptive ThinkingThinking lets Claude reason before it answers. On current Claude models it is adaptive (the model decides how much to think) and is steered with an effort level from Low to Max: in the apps through the model menu next to the send button, on the API with output_config.effort. Thinking tokens are billed as output.
-
ConnectorsClaude connectors link Claude to apps such as Gmail, Google Drive, Slack, Notion and Asana so it can read and act in them with your permissions. They are on every Claude plan, Free included; Free accounts may add one custom connector. -
Claude Desktop AppThe Claude desktop app is a free download for macOS and Windows, with Linux in beta, that adds quick entry, desktop extensions, Cowork on your own files, computer use and Claude Code to the Claude chat.