OpenAI to Give 50,000 Grants for Free ChatGPT Plus to 150,000 Scientists
On November 23, OpenAI opens applications for its free ChatGPT program for scientists: 50,000 grants, each covering a researcher plus two colleagues, meaning up to 150,000 people with Plus subscriptions. Applications are processed on a first-come, first-served basis, not by lottery. Pilot participants (a survey of 1,800 people) estimated that without AI their work would have taken almost twice as long. Labs should assemble their trio in advance and apply in the first few days.
AI-processed from OpenAI Blog; edited by Hamidun News
OpenAI is turning the summer pilot of free ChatGPT for scientists into a mass program. Applications for 50,000 grants open on Monday, November 23. Each grant is a free ChatGPT Plus subscription (a standard paid tier, not the most expensive one) for the researcher and up to two of their colleagues. That makes a team of three per grant, so up to 150,000 scientists in a single round. The program's overall goal is 200,000 researchers worldwide, twice what was planned at the July launch. Free subscriptions often look like marketing, but here the company has pilot data in hand, and it is changing the rules based on that data.
What Exactly Is Changing in the Program
Applications will be processed on a first-come, first-served basis: access goes to the first 50,000 applications that meet the criteria. In the summer, winners were chosen by lottery, but after participant feedback OpenAI dropped it. This is simpler and more predictable, but those who apply early win.
The second change concerns the tier. In the pilot, researchers were given ChatGPT Pro. When OpenAI looked at how people actually used it, it found that most would have been well served by Plus limits. The expensive tier was excessive, and the freed-up resources let the company reach more teams. That is why one grant can now be shared among three people.
Those already on the waitlist who meet the conditions will receive an email with activation instructions before applications open. Participants get access to OpenAI's most advanced models and training on using ChatGPT in research. By default, data shared through the program is not used for model training.
What the Pilot Showed
Since July, 20,000 researchers from 33 countries have joined the program. OpenAI cites two examples. Shuqi Li of Brigham and Women's Hospital and Harvard Medical School is looking for new drug targets that could shorten tuberculosis treatment. He is testing whether ChatGPT can link laboratory data with patient data and speed up antibiotic development. Arthur Porto of the Florida Museum of Natural History, together with his team, is turning photographs of museum specimens into searchable databases. They also built a 3D model of a butterfly whose wings change color when rotated, like the real specimen.
Alongside the announcement, OpenAI economists published the report Science at the Frontier: How Scientists Use ChatGPT. It analyzes how more than 1,800 researchers used ChatGPT over the summer. Participants themselves estimated that without AI they would have done the same work, at the same quality, in almost twice the time. A caveat is needed here: this is self-assessment in a survey, not an independent measurement. Moreover, the respondents were people who chose to join the program and are likely to view AI favorably.
Other Access Channels
The academic program is just one of the routes. OpenAI names three more:
- ChatGPT Edu: universities open access to the whole campus for research and teaching, with data analysis tools and administrative controls.
- GPT-Rosalind: a model for biologists with specialized reasoning and access to scientific tools and databases. Through the Rosalind Academic Plan, eligible labs at US universities receive up to 30 free seats for up to 12 months. Access requires approval through a trusted access process.
- For mathematicians, OpenAI contributed $5 million to a new initiative by the Simons Foundation. Its goal is to help specialists verify, understand, and build on results produced by AI.
What This Means in Practice
If you run a lab or work at a university, the rule is simple: apply right after November 23, since selection runs first-come, first-served. Check the criteria on the program page and decide in advance which two colleagues will join your grant. Those who already have Pro won't need to switch to Plus, but it's worth reviewing spending in light of OpenAI's own finding: most scientists found Plus limits sufficient.
For the industry, this is also a bet on habit. Researchers who learn ChatGPT in the lab tend to choose it later for projects, grants, and startups. Competitors such as Anthropic and Google will likely respond with similar programs. But independent data on how much AI actually accelerates science is still scarce. OpenAI's report offers a first picture, and it will have to be tested against real results: publications, discoveries, and reproducibility.
- Nov 21, 2026Apple Is Building an AI Agent and Making It Harder for Outsiders to Get Access: Permission Now Takes 14 Steps, Not 4
- Nov 21, 2026Developer Skips Pinecone: Searching 72,450 Code Chunks in 50 ms
- Nov 21, 2026Gemini 4 Neon shrank the TS Rust binary from 47.44 to 7.97 MB in under 10 hours
- Nov 21, 2026FDE job postings grew 4.6x in a year: what a forward deployed engineer actually does at the client's site
Key Questions Answered
What Exactly Is Changing in the Program?
Applications will be processed on a first-come, first-served basis: access goes to the first 50,000 applications that meet the criteria. In the summer, winners were chosen by lottery, but after participant feedback OpenAI dropped it. This is simpler and more predictable, but those who apply early win. The second change concerns the tier. In the pilot, researchers were given ChatGPT Pro. When OpenAI looked at how people actually used it, it found that most would have been well served by Plus limits.
What the Pilot Showed?
Since July, 20,000 researchers from 33 countries have joined the program. OpenAI cites two examples. Shuqi Li of Brigham and Women's Hospital and Harvard Medical School is looking for new drug targets that could shorten tuberculosis treatment. He is testing whether ChatGPT can link laboratory data with patient data and speed up antibiotic development. Arthur Porto of the Florida Museum of Natural History, together with his team, is turning photographs of museum specimens into searchable databases.
What are the key details on Other Access Channels?
The academic program is just one of the routes. OpenAI names three more: ChatGPT Edu: universities open access to the whole campus for research and teaching, with data analysis tools and administrative controls. GPT-Rosalind: a model for biologists with specialized reasoning and access to scientific tools and databases. Through the Rosalind Academic Plan, eligible labs at US universities receive up to 30 free seats for up to 12 months. Access requires approval through a trusted access process.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.