Independent · Source-cited · No sponsored rankingsSources linked in every article
AIToolRanked
AI News · 5 min read

ChatGPT for Academic Researchers: 50,000 free Plus grants

Applications for 50,000 free ChatGPT Plus grants open on November 23, 2026. Faculty and postdocs who qualify are accepted first come, first served.

RA
· Founder, AIToolRanked
A laptop with an abstract chat and chart, a molecule diagram, a notebook with a line chart, and three person avatars linked to the laptop
On this page

What OpenAI announced on November 16

On November 16, 2026, OpenAI published Broadening access to AI for scientists ↗. The seat rules are on the Help Center page ↗. The Next Web ↗ and Semafor ↗ reported the expansion. The picture above is ours. We have not applied for a grant, and we have not used a researcher seat.

Applications for the next round open Monday, November 23, 2026. OpenAI will award 50,000 new grants to eligible applicants, first come, first served, in place of the lottery used in the July pilot. Each grant gives the recipient and up to two collaborators free ChatGPT Plus. The Help Center states the current offer as three complimentary Plus seats for up to 12 months. Collaborators come from the same institution. A collaborator who joins later receives the remaining period. Joining later does not restart the term.

This round supports up to 150,000 researchers. The goal for the whole program is 200,000 researchers, double the goal OpenAI set when the pilot launched in July.

Who qualifies, and what changed from the pilot

The July pilot gave ChatGPT Pro. OpenAI found that most participants would have stayed within Plus limits, so new grants are Plus and cover more teams. The Next Web ↗ reports that in July each researcher was allowed to invite up to four collaborators. The new grant cuts that to two.

Since July, 20,000 researchers in 33 countries have joined. OpenAI's Economic Research team studied more than 1,800 of them. Surveyed participants estimated that the same work, at the same quality, would have taken nearly twice as long without AI. The Next Web says the estimate is self-reported, and that OpenAI says the study does not prove AI speeds up discovery. It also reports the split of science messages: reading and learning 28 percent, writing 27 percent.

The Help Center lists three tests: research faculty or a postdoctoral researcher at an eligible academic institution in a supported country; research published in the past three years in an eligible field; and a plan to use ChatGPT for academic research. The fields are Biological Sciences, Chemistry and Materials, Computer Science, Earth and Planetary Sciences, Engineering, Mathematics, and Physics. Meeting the list does not guarantee approval. Students are not on the applicant list. Eligible collaborators do include research students at the same institution, and a student collaborator cannot apply independently as the owner.

New applications are closed as of this note. The waitlist is informational and is not an application. OpenAI says people already on the waitlist who qualify get an email before applications open, with instructions for activating the account.

The workspace uses business-level protections. Its content is not used to train OpenAI's models by default. An admin can opt in to sharing it for model improvement. The offer covers ChatGPT and does not include API credits. This note states no Plus price. The announcement prints none.

Other routes on the same post: ChatGPT Edu, for a university buying campus access; the GPT-Rosalind Academic Plan, up to 30 free seats for up to 12 months for eligible U.S. university labs, after trusted-access approval; and $5 million to a Simons Foundation initiative for mathematicians examining AI-produced results. The Next Web ↗ names two rivals. Anthropic launched a Claude Science workbench in June. Google DeepMind opened science models to 200 European researchers in October.

What people are saying

  • "Plus instead of Pro is a step down from the pilot, but free frontier access with data not used for training by default is still a genuinely good deal for researchers" — fengshao li ↗ (@FengshaoL)

  • "First come, first served is a funny way to pick scientists." — DailyAI ↗ (@DailyAIOfficial)

  • "If most researchers stay inside Plus limits, the saving is real. The cost to watch is support and training, not seats." — Michael Du ↗ (@mhrdu001)

Our take

If you qualify, the useful work happens before November 23. First come, first served is the selection rule. Have the institutional email ready, a publication from the past three years in one of the seven fields, and the names of up to two collaborators at the same institution. File when the window opens. A queue hands out seats in order. It does not rank the science. That is DailyAI's objection.

Plus is a step down from the pilot's Pro seat, and the collaborator count drops from four to two. OpenAI's own review is the reason for the tier: most pilot researchers would have stayed inside Plus limits. Michael Du's split follows from that review. If usage stays inside Plus, the seat saving is real, and the cost to watch is support and training. The announcement says participants get training on using ChatGPT in research. The Help Center says the seats do not include API credits. The tier is lower than the pilot. The default, that workspace content is not used for training, is what still makes the grant worth filing.

Faculty and postdocs are who can file. A research student at the same institution can sit in a collaborator seat on someone else's grant. That seat is not an application of their own. A campus that already runs ChatGPT Edu should start with the university admin, not with this form. Literature review is a different product, in the Elicit review. Free tiers of Claude and Gemini are in ChatGPT vs Claude vs Gemini and the ChatGPT alternatives guide. Students looking for a coding tool, not this research grant, should read free AI coding tools for students.

Watch the email, the clock, and the data setting. Qualifying waitlisters are due a message before the window opens. The researcher workspace is separate from a personal ChatGPT account. Training use of workspace content stays off unless an admin opts in. GPT-6 Astra in the science-blog note is a physics result, not this grant. ChatGPT for Healthcare is a clinician account. Neither one submits the November 23 application.

More from the news desk · About Rai Ansar