OpenAI announced on November 16, 2026 that it will expand ChatGPT for Academic Researchers toward a goal of 200,000 researchers worldwide. Applications for 50,000 new grants open November 23. For university research teams, the important changes are first-come selection, Plus access for new awards and a smaller collaborator allowance than the summer pilot.
What OpenAI announced
According to OpenAI’s announcement, each new grant will cover its recipient and up to two collaborators. The company says this round can therefore support up to 150,000 researchers. The 200,000 figure is its goal for the whole programme, rather than the number of new grants or researchers already enrolled.
Semafor reported the expansion on November 16, highlighting the 50,000 grants and the doubling of the programme’s original target. OpenAI says 20,000 researchers in 33 countries have joined since the pilot began in July.
The application round opens Monday, November 23. OpenAI says eligible people already on its waitlist will receive an email before applications open explaining how to activate their accounts. Its announcement does not give a clock time for the opening.
New grants change the access and selection terms
The next round replaces lottery-based selection with access for the first 50,000 applicants who meet eligibility requirements, OpenAI says. It also changes the plan offered: the summer pilot provided Pro access, while new grants provide Plus. OpenAI attributes that decision to its review of pilot usage, saying most participants would have remained within Plus usage limits.
The company’s researcher access guide identifies the next round as three Plus seats for selected applicants, with a duration of 12 months. It lists research faculty and postdoctoral researchers as applicants and describes availability as all regions except China and Hong Kong. Identity and academic eligibility must be verified, and the guide says final eligibility and collaborator rules for the new intake remain to be confirmed.
That guide also specifies that API credits and specialized models are not included in these grants. Earlier Pro awards and grants for the existing waitlist have separate terms. Research teams should therefore use the conditions for their own intake when planning access, rather than assume that descriptions of the summer pilot apply unchanged.
What the accompanying study actually measures
Alongside the expansion, OpenAI released Science at the Frontier, its analysis of how participants used ChatGPT. The Economic Research team examined 580,186 scientific messages from 1,819 scientists between August 20 and September 22, 2026. Its survey ran from September 30 to October 7 and presents responses from 251 scientists.
Participants estimated that completing the same work at the same quality without AI would have taken about 1.9 times as long. OpenAI calculated that average across eight research activities, giving each activity equal weight. When a respondent reported no AI use for an activity, the calculation treated the time as unchanged.
The report also says 98% of the 250 scientists who answered its discovery-timeline question expected AI to accelerate discovery in their field. These are expectations from programme participants. OpenAI explicitly says the findings do not establish causal effects: productivity is self-reported, and the discovery figures describe anticipated outcomes.
That distinction matters for university evaluations of the offer. The study documents use and participants’ perceptions, while leaving the question of measured changes in research outcomes open. It also identifies staffing, funding and experimental capacity as constraints. Free model access does not itself demonstrate that those other limits have been removed.
Background and the research-office implications
Axios reported the original 100,000-researcher goal when the programme launched on July 29. The new announcement changes that target and the terms of the next intake. It gives research offices a dated opportunity to circulate, together with the specific plan and collaborator limits.
OpenAI says data shared through this programme is not used to train its models by default. Its study nevertheless analyzed usage through automated classifiers; the research team says it saw activity categories without reading the underlying messages or responses. The training default should therefore be described precisely. It is not a statement that no usage analysis occurs.
For research offices, an appropriate implementation question is which proposed workflows fit their existing permissions and data management plans. Teams evaluating the offer can record the plan, model and workflow used in a reproducible AI experiment, and document assistance through a generative AI disclosure statement where their publication or institutional rules require one. These are practical implications of wider access, rather than additional conditions announced by OpenAI.
What happens next
The next dated milestone is the November 23 application opening. OpenAI says it will use participant feedback to refine future rounds and support researchers with training. Its access guide distinguishes this offer from institutional Edu arrangements, the Rosalind Academic Plan for eligible life-science labs and the separate Researcher Access Program for API credits. For a research team deciding how to apply, the immediate task is to identify its correct route and the terms attached to that route.
Sources
- OpenAI: Broadening access to AI for scientists, November 16, 2026. Checked 2026-11-16.
- Semafor: OpenAI announces programme boosting access to frontier AI for scientists, November 16, 2026. Checked 2026-11-16.
- OpenAI: Plans and programmes for researchers. Checked 2026-11-16.
- OpenAI: Science at the Frontier, November 16, 2026. Checked 2026-11-16.
- Axios: OpenAI offers 100,000 academics free ChatGPT access, July 29, 2026; background on the original launch. Checked 2026-11-16.







