OpenAI adds 50,000 free ChatGPT research grants, replacing the application lottery
Each grant covers a researcher and up to two collaborators. New recipients get Plus rather than Pro as OpenAI doubles its overall reach goal to 200,000 researchers.
Loading page…
Each grant covers a researcher and up to two collaborators. New recipients get Plus rather than Pro as OpenAI doubles its overall reach goal to 200,000 researchers.
Listen to this story
OpenAI’s academic researcher program is shifting from lottery awards to first-come, first-served grants, with 50,000 new awards opening November 23, 2026. The expansion raises the program’s intended worldwide reach to 200,000 researchers, though 20,000 have joined so far; each grant covers a researcher and up to two collaborators. Recipients will get ChatGPT Plus instead of the pilot’s Pro tier, based on OpenAI’s finding that most participants would have stayed within Plus limits. The reported productivity benefit remains self-estimated, not a measured causal effect.
OpenAI’s study analyzed more than 580,000 messages from 1,819 scientists and surveyed 251 participants; respondents estimated their work would take 1.9 times longer without AI.
The research team used automated message classifiers and could see activity categories, but did not read participants’ underlying messages or ChatGPT responses.
Qualifying researchers already on the waitlist will receive an email before applications open with instructions to activate their accounts.
OpenAI will award 50,000 new grants for free ChatGPT access, with applications opening November 23, 2026. The next round of ChatGPT for Academic Researchers replaces a lottery with first-come, first-served selection for eligible applicants. Each grant covers the recipient and up to two collaborators, putting the round’s potential reach at 150,000 researchers.
The expansion, announced November 16, raises the program’s overall goal to 200,000 researchers worldwide—double the target set when the pilot began in July. OpenAI says 20,000 researchers in 33 countries have joined so far. The new target describes the program’s intended reach, not the number already receiving access.
New grants will provide ChatGPT Plus rather than the Pro subscriptions supplied during the pilot. OpenAI says its review of pilot usage found that most participants would have stayed within Plus usage limits. It is using that finding to offer a different tier while extending access to more research teams.
OpenAI says researcher feedback prompted it to move away from lottery-based awards. For this round, access will go to the first 50,000 applications that meet the eligibility requirements.
Alongside the expansion, OpenAI published Science at the Frontier, a study of how participants used ChatGPT. Its Economic Research team examined more than 580,000 scientific messages from 1,819 scientists and surveyed 251 participants. The work covered activities including reviewing literature, analyzing evidence, writing code and developing research ideas.
To analyze usage while protecting privacy, OpenAI used automated classifiers—tools that sort messages into categories such as literature review or data analysis. The research team could see those categories but did not read the underlying messages or ChatGPT’s responses. The analysis gave each scientist equal weight when averaging the share of messages devoted to different activities.
The expansion announcement also describes work underway among participants. Shuqi Li, a researcher at Brigham and Women’s Hospital and Harvard Medical School, is exploring whether ChatGPT can link laboratory and patient data to speed antibiotic development. At the Florida Museum of Natural History, Arthur Porto and his team use AI tools to turn specimen photographs into searchable databases.
Participants can use OpenAI’s frontier models and receive training on using ChatGPT in research. OpenAI says data shared through the program is not used to train its models by default. It also plans to use participant feedback on the application process, needed capabilities and support to refine future rounds, while working with scientific communities on field-specific needs.
Researchers already on the waitlist have a separate immediate step: OpenAI says qualifying applicants will receive an email before applications open, with instructions for activating their accounts.
Story updates
OpenAI’s analysis of its summer academic pilot found that researchers used ChatGPT across scientific work, while the study’s time-saving and discovery findings are based on participants’ estimates—not measured research outcomes.
Scientists in OpenAI’s academic pilot used ChatGPT for more than writing and coding: they also brought it into developing research ideas. Findings published November 16 show broad uses and large perceived benefits. Surveyed participants estimated that the same work, at the same quality, would have taken 1.9 times as long without AI.
OpenAI’s Economic Research team analyzed more than 580,000 scientific messages from 1,819 scientists in the ChatGPT for Academic Researchers summer pilot. It also surveyed 251 participants. The study combines a record of what researchers asked the chatbot to do with their own assessments of its value—two different kinds of evidence about AI’s place in scientific work.
The team used automated classifiers—tools that sort messages into categories—to identify activities such as literature review, coding and data analysis. Researchers could see those categories but did not read the underlying messages or ChatGPT’s responses. For its activity breakdown, the team calculated each scientist’s share of messages in each category, then averaged those shares so every scientist received equal weight.
OpenAI found that 68% of the scientists used ChatGPT to develop research ideas. Its breakdown of ideation messages distinguishes between researchers bringing a specific question, asking for help refining an existing idea, and requesting a question from the model. That makes the headline use case more specific than simply asking AI to invent a project.
The remaining 6% of ideation messages fell outside those three categories. Scientists also stayed involved after delegating tasks: OpenAI tracked follow-up requests to clarify, check or revise responses across research activities. Separately, the company says data shared through the academic program is not used to train its models by default.
The survey asked scientists to imagine a discovery that would take 10 years without AI and estimate its timeline with AI. Of the 250 participants who answered, 98% expected an acceleration in their field. Respondents could select multiple fields, so the field-level groups overlap. The result captures expectations about future discovery, rather than elapsed time to an actual scientific result.
Likewise, the estimate about completing comparable work without AI came from participants themselves. Maintaining the same quality was part of what they were asked to estimate. It should be read as a reported benefit, not as a measured reduction in the time needed to produce a discovery.
Asked why they expected faster discovery, scientists pointed beyond efficiency alone. OpenAI identified 367 reasons in 224 written responses, with respondents able to give more than one reason. The three leading categories accounted for the following shares of reasons, not shares of all scientists:
Loading discussion...
52% concerned getting research done faster.
23% concerned finding more promising research questions.
14% concerned pursuing a wider range of questions.
OpenAI acknowledges that establishing whether those expectations translate into faster discovery will require more research. It also names constraints outside the chatbot: scientists still need funding, staff and experimental capacity to test their ideas. The study’s account of digital assistance does not remove those requirements from the path to a result.
Alongside the findings, OpenAI set out an expansion of ChatGPT for Academic Researchers. Its program announcement gives an overall goal of 200,000 researchers worldwide, double the target at the July pilot’s launch. That is an intended reach, not current enrollment: OpenAI says 20,000 researchers in 33 countries have joined since the pilot began.
The pilot supplied ChatGPT Pro, but new grants will provide Plus. OpenAI says most pilot participants would have stayed within Plus usage limits, allowing it to offer access to more teams at that tier. Each new grant covers a recipient and up to two collaborators, giving the next round a potential reach of 150,000 researchers.
Applications open November 23, 2026, for 50,000 new grants. Following researcher feedback, OpenAI is replacing lottery-based selection with first-come, first-served awards for eligible applicants. Qualified researchers already on the waitlist will receive an email before applications open explaining how to activate their accounts.
Join the conversation
Explain which approach gives researchers a fairer chance.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.