ChatGPT deep research limits are set per plan, but OpenAI no longer publishes a task count for Free, Go, Plus or Pro. Its pricing page calls the allowance limited on Free and Go, expanded on Plus and maximum on Pro, and the counter in ChatGPT shows how many tasks you have left. Fixed monthly allowances reset 30 days after your first use. Business, Enterprise and Edu workspaces can go past their included use with credits, at about 50 credits per task.
Last checked November 9, 2026, against OpenAI’s Deep research in ChatGPT help article, the ChatGPT pricing page and the ChatGPT rate card for credit-based pricing.
On this page
Deep research limits by plan
OpenAI describes the personal plans’ allowance in words, not numbers. The table quotes those words, and for work plans it gives what OpenAI’s credit articles say.
| Plan | What OpenAI publishes | Apps for deep research | When the allowance runs out |
|---|---|---|---|
| Free | “Limited” | Limited | Wait for the reset |
| Go | “Limited” | Limited | Wait for the reset |
| Plus | “Expanded” | Yes | Wait for the reset |
| Pro | “Maximum” | Yes | Wait for the reset |
| Business | Per-seat limits, no count published | Depends on workspace apps | Workspace credits, if the owner bought them |
| Enterprise | Shared credit pool, no per-seat cap by default | Depends on workspace apps | Credits from the pool |
| Edu | 5 queries per rolling 24 hours included | Depends on workspace apps | Credits from the pool |
Older articles still quote fixed numbers such as 25 tasks a month on Plus. That figure came from an April 2025 update to OpenAI’s deep research announcement and does not appear on its current help or pricing pages, so the only reliable count for a personal plan is the one ChatGPT shows you.
How the counter and the 30-day reset work
ChatGPT counts deep research in tasks. A task is one research run: you describe the outcome, ChatGPT proposes a research plan you can edit, and you get a cited report at the end. OpenAI’s help article says your in-product usage counter shows your remaining tasks.
For plans with a fixed monthly allowance, the count resets every 30 days from the date of your first use, not on the first of the month. If you first ran deep research on the 12th, your allowance renews around the 12th of the next cycle.
Availability also depends on your country or territory, so a plan that lists deep research may not offer it everywhere.
Business, Enterprise and Edu: credits
Work plans add a second meter. OpenAI’s flexible pricing article says Business users get per-seat limits for advanced features, deep research among them, and can keep going from a shared credit pool once a limit is reached, if the workspace has bought credits. Enterprise and Edu workspaces buy a shared pool at the contract level, with no per-seat caps by default; admins can set spend controls by group.
The rate card prices one deep research task at about 50 credits. For comparison, it lists an image generation at 5 credits and an agent mode message at 30.
ChatGPT Edu includes 5 deep research queries per rolling 24-hour period at no extra cost. The window starts at your first use: run one at 10:00 and the window lasts until 10:00 the next day. Usage beyond that draws credits at the same rate.
Can Plus or Pro buy more deep research?
Not with credits, going by OpenAI’s current wording. The help article on credits for personal plans lists Codex, ChatGPT Work, Word, Excel and PowerPoint as the supported features, and the Pro tiers article says credits do not add usage for every ChatGPT model or feature. Deep research is not on that list, so on a personal plan the options are to wait for the reset or move to a plan with a larger allowance.
Ways to spend fewer tasks
- Use search or standard chat for quick facts. OpenAI suggests saving deep research for multi-step questions that combine many sources.
- Edit the research plan before the run starts instead of rerunning a task with a vague brief.
- Limit the sources: under Sites > Manage sites you can restrict research to chosen domains, or prioritize them while still searching the web.
- Interrupt a run that has drifted and adjust its focus rather than starting over.
For writing a tighter brief, the prompt chaining guide and the ChatGPT prompt for comparing documents show how to state the goal, sources and output format up front.





