top of page

ChatGPT for Academic Researchers opens 50,000 Plus grants on November 23

59 minutes ago
6 min read
ChatGPT for Academic Researchers opens 50,000 Plus grants on November 23, Data Studios original cover

OpenAI is expanding ChatGPT for Academic Researchers, the free-access program it piloted in July, with 50,000 new grants that open to applicants on November 23. The company said in a post on its website on November 16 that the program now aims to reach 200,000 researchers, twice the goal it set when the pilot launched, and that the new grants will be handed out to the first eligible applicants instead of through the selection process used so far.


The expansion comes with a change of plan. Researchers admitted during the pilot received ChatGPT Pro, OpenAI's top consumer tier, while each new grant includes ChatGPT Plus. OpenAI said most pilot participants would have stayed within Plus limits anyway, and presented the switch as the way to fund access for many more teams. Each new grant covers one researcher and up to two collaborators, so a single award can put up to three people on Plus, and the 50,000 grants together can reach up to 150,000 people.


According to OpenAI, about 20,000 researchers in 33 countries have joined the program since July. Qualifying applicants already on the waitlist will receive activation instructions by email before the new round opens to everyone else, the company said.


··········


CHATGPT FOR ACADEMIC RESEARCHERS: PILOT AND EXPANSION COMPARED.


The program launched at the end of July with a goal of 100,000 researchers through 2027 and a starting cohort of 10,000. In OpenAI's July announcement, participants received access to its frontier models, at the time the GPT-5.6 family, along with training, hands-on support and peer learning, and each participant could invite up to four collaborators. OpenAI also stated then that researcher data is not used to train its models by default, and it repeated that commitment for the expanded program.


The November changes touch almost every operational parameter of the pilot. The table sets the two phases side by side using the figures OpenAI has published.


........


Parameter

July pilot

November expansion

Plan included

ChatGPT Pro

ChatGPT Plus

Researcher goal

100,000 through 2027

200,000

Grants in this round

10,000 at launch

50,000 new grants

People per grant

Researcher + up to 4 collaborators

Researcher + up to 2 collaborators

Selection

Lottery

First eligible applicants

Applications

Program page, then waitlist

Open November 23; waitlist first

Training on data

Off by default

Off by default


........


The per-grant seat count falls from five people to three, a reduction of 40% in the maximum size of a funded team. In exchange, the number of grants in a single round is five times the pilot's starting cohort. For a principal investigator with a large lab, the new format means choosing two collaborators instead of four; for a small group or a solo researcher, the change of seat count makes no practical difference.


··········


CHATGPT PLUS VS CHATGPT PRO FOR GRANT RECIPIENTS.


The most concrete question for new applicants is what Plus offers compared with the Pro access that pilot participants received. OpenAI's official plan comparison page, checked on November 17, 2026, lists the following differences between the two tiers. The page did not display US prices for either plan when checked, so the comparison below covers capabilities and limits only.


........


Feature

ChatGPT Plus

ChatGPT Pro

GPT Instant context window

54K tokens

128K tokens

GPT Reasoning context window

256K tokens

400K tokens

Reasoning input maximum

About 320 pages

About 680 pages

GPT-5.6 Sol Pro

Not included

Included

Pro reasoning (GPT-6 Astra)

Not included

Included

Deep research

Expanded

Maximum

Codex

Expanded usage

Longer sessions

Image creation

More complex images

Unlimited and faster

Response times

Fast

Ultrafast on eligible tiers

Monthly API credits

None listed

Included (amount by tier)


........


For most reading, drafting and analysis tasks, the difference is a matter of limits. Plus includes GPT-6.1 Sol, GPT-6 Astra and GPT-5.6 Terra, together with projects, scheduled tasks, custom GPTs, expanded deep research and expanded Codex usage, which covers the uses OpenAI says dominate scientific work in the program. The gaps show up in long-document and heavy-compute workflows. A reasoning context of 256K tokens instead of 400K, and a maximum of roughly 320 pages of input instead of 680, will matter to researchers who load long theses, large literature sets or entire datasets into one conversation. Pro reasoning and GPT-5.6 Sol Pro are absent from Plus, and so are the monthly API credits that Pro subscribers receive. Our comparison of ChatGPT Plus and ChatGPT Pro covers the two plans' usage tiers in more detail, and our report on the API credits included with ChatGPT Pro and Business explains what Plus grant holders will not receive.


··········


HOW TO APPLY FOR A CHATGPT PLUS RESEARCH GRANT.


OpenAI said applications for the new round open on November 23 and that grants will go to the first applicants who meet the eligibility criteria. Researchers who joined the waitlist earlier and qualify will be contacted first by email with activation instructions. The program was designed in July for scientists, mathematicians and engineers working in academic research.


Several details that applicants will want to know have not been published in the company's statement. OpenAI has not set out the full eligibility rules, such as which institutions, career stages or fields qualify, and it has not said how long a Plus grant lasts or whether it renews. The company also has not said whether researchers already in the pilot keep their Pro access or will be moved to Plus when their current access ends. Because grants follow a first-come, first-served rule, researchers who are already eligible have a practical reason to prepare their application details before November 23 and to apply on the first day.


··········


OPENAI'S STUDY OF 1,800 SCIENTISTS IN THE CHATGPT PILOT.


Alongside the expansion, OpenAI published a study of more than 1,800 scientists who used ChatGPT during the pilot. On average, reading and learning accounted for 28% of their science-related messages and writing for 27%, meaning that more than half of the activity went to understanding literature and producing text rather than to calculation or code. Surveyed participants said the same work would have taken nearly twice as long without AI. That figure is self-reported, and OpenAI said the study does not show that AI speeds up scientific discovery itself. The company listed funding, staff and laboratory capacity as the constraints that still limit how many new ideas researchers can test.


These usage shares support OpenAI's argument for the plan switch. Literature review and writing are the tasks where Plus limits are least likely to bind, while the Pro features that Plus lacks serve longer reasoning sessions and much larger inputs. The study does not report how many pilot participants hit Plus-level ceilings, so the claim that most would have stayed within Plus limits rests on OpenAI's own usage data, which it has not released in detail. Anthropic published a comparable analysis earlier this month in its study of 207,049 Claude sessions on how scientists use AI.


··········


THE 200,000-RESEARCHER TARGET IN NUMBERS.


The published figures allow a rough check of how far the new round moves the program toward its goal. With about 20,000 researchers already enrolled, OpenAI needs roughly 180,000 more to reach 200,000. The new round can add up to 150,000 people if every grant is used with both collaborator seats filled, which would bring the total to about 170,000 and leave a gap of about 30,000. If grants are used by researchers alone, the round adds 50,000 people and the program reaches about 70,000. OpenAI has not said whether collaborators count toward the 200,000 target or whether further rounds are planned, so these totals mark the range the current round allows and are not a company forecast.


The scale is far larger than other recent academic programs from AI companies. Google has offered six months of free top-tier Antigravity access to 10,000 researchers, and Google DeepMind extended frontier AI access to 200 European researchers with $3 million in cloud credits. OpenAI's program trades depth of access for reach: a lower-tier plan for each person, distributed to many more people.


OpenAI also pointed to two related science initiatives. The Rosalind Academic Plan offers eligible laboratories at US universities up to 30 seats on GPT-Rosalind, its biology model, for up to 12 months at no cost, subject to approval. Separately, the company committed $5 million to a Simons Foundation initiative that helps mathematicians verify and build on results produced with AI.


··········


CHATGPT PLUS RESEARCH GRANTS: CONFIRMED FACTS AND OPEN QUESTIONS.


OpenAI has confirmed the main terms of the expansion: 50,000 new grants, ChatGPT Plus in place of Pro, one researcher plus up to two collaborators per grant, a first-come allocation, applications from November 23 with waitlisted applicants contacted first, a target of 200,000 researchers, and the continued exclusion of program data from model training by default. The enrollment figure of about 20,000 researchers in 33 countries and the findings of the 1,800-scientist study also come from the company.


The terms that decide how valuable a grant is for a given lab remain undisclosed: the eligibility rules, the length of a Plus grant, the treatment of existing Pro participants, and whether collaborators count toward the researcher goal. On the plan itself, the official comparison page shows that Plus covers the reading, learning and writing work that made up most pilot activity, while researchers who depend on 400K-token reasoning sessions, inputs of several hundred pages or Pro-only models will find the new grants narrower than the access the first cohort received.


··········


FOLLOW US FOR MORE.


·····


DATA STUDIOS


·····


datastudios.org


bottom of page