ChatGPT deep research made the category famous in February 2025. Its 2026 version is less about raw depth and more about control: which sites it reads, which apps it can open, and when you step in.
Key Takeaways
- On GPT-5.2 since February 2026. — The update added site-restricted research, app and MCP connections, steering mid-run and a better report view.
- Access by label, not count. — Limited on Free and Go, expanded on Plus, maximum on Pro, included on Business and Enterprise. Fixed allowances reset 30 days after first use.
- Also an API. — o3-deep-research and o4-mini-deep-research run in the OpenAI API with MCP and webhooks.
What does ChatGPT deep research do?
Deep research takes a question, asks clarifying questions, then searches, reads and analyzes sources for 5 to 30 minutes before returning a cited report (OpenAI). OpenAI's own limitations note says it can hallucinate facts, make incorrect inferences and struggle to separate authoritative information from rumor.
The February 2026 update moved it to a GPT-5.2-based model and added three things buyers ask for: research limited to specific sites, connections to apps and MCP servers, and a better report interface (Wikipedia). Users can interrupt to steer or add sources. Developers get o3-deep-research and o4-mini-deep-research in the API with MCP and webhooks (The Decoder).
How does ChatGPT deep research score?
Editorial scores from published documentation and announcements through October 31, 2026, not benchmark results. Total: 17/20, on the same criteria as the Best AI Research Tools guide.
| Criterion | Score | Why |
|---|---|---|
| Depth | 5/5 | 5 to 30 minute runs, steerable mid-task, with a mature report view. |
| Citations | 4/5 | Inline citations and site restriction; OpenAI says it can still hallucinate and confuse rumor with authority. |
| Reach | 4/5 | Web, uploaded files, connected apps and MCP servers. |
| Access | 4/5 | Limited free access, Plus at $20 a month, plus an API. |
How is ChatGPT deep research priced?
It comes with ChatGPT plans, not as a separate product. As published in October 2026: limited on Free and Go ($8 a month), expanded on Plus ($20), maximum on Pro (from $100), and included on Business ($25 per seat, or $20 billed annually) and Enterprise, with credits for use beyond the included amount (ToolColumn, citing OpenAI). API use is billed per token and per search call.
Who is ChatGPT deep research for?
- Fits: general web research for people already on ChatGPT; teams that want site-restricted research or MCP-connected internal sources.
- Look elsewhere: if you want the widest free access (Gemini Deep Research), research over Google Workspace mail and docs (Claude Research), or scientific literature (Elicit).
What are the ChatGPT deep research alternatives?
- Gemini Deep Research: available on every Gemini plan, plus Deep Research Max in the API.
- Perplexity Deep Research: faster, citation-first reports.
- Microsoft 365 Copilot Researcher: research over Microsoft 365 work data.
- Compare: ChatGPT vs Gemini vs Perplexity deep research.
Related questions
- Which AI research agents are worth using in 2026?
- ChatGPT, Gemini or Perplexity: which deep research should you use?
- How reliable are research agent citations, and how should a company govern them? (partner site CTAIO)
- How was deep research trained to browse?
Sources
- OpenAI, Introducing deep research (Feb 2, 2025, updated Feb 10, 2026).
- Wikipedia, ChatGPT Deep Research.
- ToolColumn, ChatGPT deep research limits by plan (Oct 11, 2026, citing OpenAI pricing and help pages).
- The Decoder, OpenAI expands API access to deep research models (Jun 26, 2025).
How many deep research tasks do I get in ChatGPT?
OpenAI no longer publishes fixed counts for personal plans. As of October 2026, deep research is labelled limited on Free and Go, expanded on Plus and maximum on Pro; Business and Enterprise include it with credit-based overage. ChatGPT shows a counter of remaining tasks in the product.
What model does ChatGPT deep research use?
It launched in February 2025 on a version of o3. Since February 2026 it runs on a GPT-5.2-based model. Earlier overflow queries used a lighter o4-mini version.
Is ChatGPT deep research available in the API?
Yes. OpenAI added o3-deep-research and o4-mini-deep-research to its API in June 2025, with web search, code execution, MCP support and webhooks that notify you when a long task finishes.
Ready to Find the Right AI Tools?
Browse our data-driven rankings to find the best AI tools for your team.