Gemini Deep Research
ChatGPT deep research
The easiest to get, on every Gemini plan including free, with the most developer-ready API agents and private-data search through MCP
The most control over what it reads: site restriction, connected apps and MCP, and steering mid-run
- Free in the app
- Deep Research Max
- MCP and File Search
- AI Pro $19.99
- Site restriction
- Apps and MCP
- Mature reports
- API in preview
- Max not in the app
- Limits unpublished
- Long runs
Key Takeaways
- Gemini Deep Research — Listed on every Gemini plan including free. April 2026 API agents on Gemini 3.1 Pro: Deep Research and Deep Research Max, with MCP, File Search and charts. 18/20.
- ChatGPT deep research — GPT-5.2-based since February 2026, with site restriction, app and MCP connections and mid-run steering. Limited on Free and Go, Plus $20. 17/20.
- Perplexity Deep Research — Fastest of the three, citation-first. Advanced Deep Research adds clarifying questions, a code sandbox and editable reports. Pro $20. 15/20.
- Bottom line — Gemini for free access and building on an API, ChatGPT for the most control over sources, Perplexity for quick cited reports inside your search habit.
What is the difference between ChatGPT, Gemini and Perplexity deep research?
All three take a question, plan searches, read sources and return a cited report. They differ in how long they run, how much control you get over sources, and how they are packaged. Gemini is bundled into every Gemini plan and exposed to developers as two agents. ChatGPT added source restriction and connections in its February 2026 update. Perplexity is the fastest and sits inside a search product.
Facts verified against vendor pages, help centers and coverage through October 31, 2026; scores match the Best AI Research Tools guide. If you want to be cited by these agents rather than use them, see agentic search.
How do the features compare?
Feature matrix
| Feature | ChatGPT deep research | Gemini Deep Research | Perplexity Deep Research |
|---|---|---|---|
| Research | |||
| Model (Oct 2026) | GPT-5.2-based (since Feb 2026) | Gemini 3.1 Pro for API agents | Claude Opus 4.6 Thinking on Max; lighter model on Pro |
| Typical run time | 5 to 30 minutes (OpenAI) | Several minutes; Max runs longer in the background | About 2 to 4 minutes at launch |
| Steering | Clarifying questions; interrupt and add sources mid-run | Collaborative plan review in the API | Clarifying questions; follow-ups mid-run |
| Sources and citations | |||
| Restrict to specific sites | Yes (Feb 2026) | API can turn off web and search only your data | Not documented for Deep Research |
| Private data | Files, connected apps, MCP servers | API: MCP servers and File Search | Document uploads |
| Vendor's stated limits | Can hallucinate; may confuse rumor with authority | Vendor benchmarks only | Benchmarks claimed, none named |
| Buying (Oct 2026) | |||
| Free access | Limited on Free and Go | On the free plan within usage limits | Limited on Free |
| Paid entry | Plus $20/mo; Pro from $100/mo | AI Plus $7.99/mo; AI Pro $19.99/mo | Pro $20/mo; Max $200/mo |
| API | o3-deep-research, o4-mini-deep-research | Deep Research, Deep Research Max (preview) | Sonar Deep Research |
| WTF score (/20) | 17 | 18 | 15 |
ChatGPT deep research
Launched February 2, 2025 and moved to a GPT-5.2-based model in February 2026, when OpenAI added research limited to specific sites, app and MCP connections and a new report view. OpenAI says runs take 5 to 30 minutes and that the agent can hallucinate and struggle to separate authoritative sources from rumor. Access is labelled limited on Free and Go, expanded on Plus and maximum on Pro.
ChatGPT deep research
- Most control over sources: site restriction, app and MCP connections, mid-run steering
- Mature long-report format with export
- API models with MCP and webhooks
- No published task counts on personal plans; heavy use needs Pro
- Runs can take up to 30 minutes
Gemini Deep Research
Google shipped Deep Research in the Gemini app in December 2024 and lists it on every plan, from free to AI Ultra. On April 21, 2026 it announced Deep Research and Deep Research Max for developers on Gemini 3.1 Pro, with MCP servers, File Search, code execution and native charts. Google reported Max at 93.3% on its DeepSearchQA benchmark.
Gemini Deep Research
- Available on every Gemini plan, including free
- Deep Research Max in the API for the most thorough background runs
- MCP and File Search for private data; Google working with FactSet, S&P and PitchBook on MCP servers
- API agents still in preview, through the Interactions API only
- Deep Research Max not announced for the Gemini app
Perplexity Deep Research
Launched February 14, 2025 with free daily queries and reports in about two to four minutes. Advanced Deep Research adds clarifying questions, follow-ups during a run, a code sandbox, document uploads and reports that stream into an editable file; Max subscribers get Claude Opus 4.6 Thinking. Pro was $20 a month and Max $200 in April 2026.
Perplexity Deep Research
- Fastest route to a cited report
- Free tier and a search-first interface many people already use
- Sonar Deep Research API priced by tokens
- Usage limits changed during 2026; check current caps
- Fewer documented private-data connections than the other two
Related questions
- Which AI research agents are worth using in 2026?
- How reliable are research agent citations? (partner site CTAIO)
- Should you use Claude Research instead?
- Which research agent is best for scientific papers?
- How do you get your pages cited by research agents?
Frequently asked questions
Is ChatGPT deep research better than Gemini Deep Research?
They are close. ChatGPT gives more control over sources: you can restrict research to specific sites, connect apps and MCP servers, and steer mid-run. Gemini is easier to get, since Deep Research is on every Gemini plan including free, and Google's Deep Research Max agent in the API targets the most thorough runs. We score Gemini 18/20 and ChatGPT 17/20 on published capabilities.
Is Perplexity Deep Research as good as ChatGPT?
It is faster and lighter. Perplexity reports took about two to four minutes at launch against OpenAI's five to thirty, and its Advanced Deep Research update added clarifying questions, a code sandbox and editable reports. For the longest, most steerable runs, ChatGPT goes further.
Which deep research tool is free?
All three have some free access. Gemini lists Deep Research on its free plan within usage limits; ChatGPT offers limited deep research on Free and Go; Perplexity's free plan includes limited research.
Which one has the most accurate citations?
No independent test has compared the current versions directly. An audit published in September 2025 (DeepTRACE) found citation accuracy between 40 and 80 percent across deep research systems. Check citations yourself for anything that matters, whichever tool you use.
Final verdict
- Choose Gemini Deep Research if you want deep research on a free or low-cost plan, or plan to build research into a product with private data. Gemini Deep Research review.
- Choose ChatGPT deep research if you need to restrict sources, connect internal apps through MCP, or steer long runs. ChatGPT deep research review.
- Choose Perplexity Deep Research if you want quick cited reports inside the search tool you already use. Perplexity Deep Research review.
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); The Decoder, Deep research models in the API (Jun 26, 2025).
- VentureBeat, Google's new Deep Research and Deep Research Max agents (Apr 21, 2026); Google AI for Developers, Gemini Deep Research agent; 9to5Google, Google AI plan features (May 25, 2026).
- Engadget, Perplexity Deep Research launch (Feb 15, 2025); Perplexity Help Center, What's new in Advanced Deep Research; Finout, Perplexity pricing in 2026 (Apr 27, 2026).
- Venkit et al., DeepTRACE, arXiv 2509.04499 (Sep 2, 2025).
Ready to Find the Right AI Tools?
Browse our data-driven rankings to find the best AI tools for your team.