ChatGPT Deep Research Limit: Runs per Plan and What to Do (2026)

The ChatGPT deep research limit depends on your plan, and OpenAI's own pricing page doesn't give a number for it. It labels deep research "Limited" on Free, "Expanded" on Plus and "Maximum" on Pro, and ChatGPT shows a counter of the runs you have left. The last fixed figures OpenAI published, in its June 2025 FAQ, were 25 a month on Plus (10 full runs and 15 lightweight ones) and 250 on Pro, and those can't be assumed to hold today. The reliable answer is the counter in your own account. Open a chat, choose deep research, and read it.

If the counter is the problem, because you run out before the month does, you have two ways out: pay for a bigger ChatGPT plan, or move the research somewhere that tells you the number up front. Zemith Plus is $15.99 a month and deep research there costs at most 100,000 credits a run against 1.5 million credits a month, so you can plan around a number instead of a label. Professional is $29.99 with 3 million.

TL;DR

• As of October 2026, OpenAI's pricing page describes ChatGPT deep research as Limited (Free), Expanded (Plus) and Maximum (Pro) and prints no run count. The in-product counter is the source of truth.
• OpenAI's last published allowances (June 2025 FAQ, as summarized by Wikipedia) were 25 a month on Plus, 10 of them full runs, and 250 on Pro. Treat them as history.
• Google's Gemini Apps limits page gives multipliers, not counts: Google AI Plus 2x, AI Pro 4x the standard limit, as of October 2026.
• Claude Pro ($20 a month, $17 billed yearly) includes Research and the free plan doesn't. Anthropic doesn't print a count either.
• Zemith charges at most 100,000 credits per deep research run. Plus includes 1.5 million credits a month, which covers up to 15 runs if you do nothing else, and Professional's 3 million covers up to 30.

What does each plan actually give you?

Most "limit" articles quote a number with no date on it. These are the figures that come from the vendors' own pages, with the gaps left as gaps.

Deep research allowances as published by each vendor, October 2026.

Sources: Zemith pricing and the credit rate in Zemith's tool list; OpenAI's pricing page and deep research help article, both of which returned errors to our fetch, so the ChatGPT labels come from the Wikipedia page above, which cites them; Google's Gemini Apps limits; Claude's pricing page. ChatGPT Plus and Google AI Pro prices are the figures we confirmed for our ChatGPT Go vs Plus and Google AI Plus vs Pro posts.

Two things stand out. Nobody except Zemith gives you the number before you start, and ChatGPT Pro still wins on raw volume: 250 runs a month in the last published figures is more than any $15.99 plan can match.

How do you check how many ChatGPT deep research runs you have left?

Start a deep research request in ChatGPT and look at the counter that appears where you pick the tool. According to OpenAI's help center, as summarized by Wikipedia, fixed monthly allowances reset 30 days after your first use, not on the first of the month. OpenAI has also said that when the full-strength allowance is used up, requests drop to a lighter version.

That fallback is the part that surprises people. You can still press the button after "hitting the limit", but the report is shorter and reads fewer sources. If the output suddenly feels thin, the counter is worth a look before you blame the prompt. A tighter prompt also stretches a run further, and our deep research prompt template is built for that.

Is it worth upgrading ChatGPT just for more deep research?

Only if research is most of what you use ChatGPT for. Plus costs $20, and on the last published numbers its 25 included runs only 10 at full strength. Moving up to a Pro tier to get "Maximum" is a big step if the reason is a handful of extra reports a month.

Before paying more, work out the real cost per report. Ten full runs on a $20 plan is $2 a report if you do nothing else with ChatGPT. That is the number to beat.

How does Zemith handle deep research differently?

You ask for deep research in a chat. Zemith drafts a plan first and waits; nothing runs until you press Start Research, so you can narrow the scope before any credits go on searching. After that it keeps running if you close the app, and the default depth reads up to 24 pages in up to 10 minutes before writing a cited report.

The price is a ceiling you can see. A run costs at most 100,000 credits, and Zemith reserves that ceiling up front and charges what the run actually used, never more than the ceiling. Here is what that means per plan, using the prices on the pricing page:

Those are the most a run can cost, not an average, and the same credits pay for everything else you do in the app. If you burn through them, a top-up is $10 for 800,000 credits, which is eight more runs at the ceiling.

One concrete week: a consultant has three client briefings. On Monday they ask for a deep research report on each client's market, review each plan, and press Start Research. The reports finish while they're in meetings, with citations, and cost at most 300,000 credits, a fifth of a Plus month. The same Plus plan also gives them Claude, Gemini and GPT in the same chat for follow-up questions, plus Word, PowerPoint and Excel files, so the report can become the deck without leaving the app. The deep research page and the researchers page show the workflow, and the Perplexity comparison covers that tool.

What about the whole stack?

A lot of people who hit the limit also pay for a second assistant. ChatGPT Plus and Claude Pro together are $40 a month, or $480 a year at monthly prices. Zemith Plus is $179.88 a year on the yearly plan and puts both families of model in one app. That is a difference of about $300 a year, and it only counts if you would use Zemith for the work you currently split across the two.

Where Zemith is the wrong pick

Don't switch if your volume is high. A researcher who needs 100 reports a month will be better served by a top ChatGPT tier than by 15 to 30 runs a month. Zemith's depth is fixed (up to about 24 pages and 10 minutes), so very long investigations that read hundreds of sources are better in a tool built for that. Deep research isn't available inside Zemith Cloud tasks, so a scheduled weekly research report isn't possible there. And if you only ever use ChatGPT and your counter lasts the month, you have nothing to fix.

How we compared

On October 26, 2026 we read Google's Gemini Apps limits page and Claude's pricing page, and tried OpenAI's pricing page and deep research help article, which refused our fetch, so ChatGPT's labels and the June 2025 figures come from Wikipedia's summary of them. We didn't use third-party counts that disagreed with each other. Zemith's credit rate, plan prices, credit allowances and the plan-first flow were checked in the product's code and plan list. We compared published prices and documentation and didn't benchmark report quality.

The bottom line

If ChatGPT's counter runs out before your month does, find out what a run really costs you before you jump to a bigger plan. If you run a few reports a week, Zemith Plus at $15.99 gives a number you can plan around, up to 15 cited runs a month, and Claude, Gemini and file creation in the same app.