OpenAI offers free ChatGPT access to 100,000 researchers

OpenAI will offer free access to its most advanced models to 100,000 researchers at universities and academic institutions through 2027. The program, called ChatGPT for Academic Researchers, is designed to help scientists, mathematicians and engineers use AI to speed up tasks ranging from reviewing literature to analyzing data and testing hypotheses.

The rollout begins this summer with 10,000 researchers from selected institutions, including the Institute for Advanced Study and the École normale supérieure. Participants will be able to invite up to four collaborators from their institution, provided those collaborators also verify their affiliation.

What the program includes

Researchers will get free access to ChatGPT, ChatGPT Work and Codex, OpenAI's tool for programming and running technical tasks. They will also receive higher usage limits, larger context windows and extended access to the deep research feature.

The GPT-5.6 family of models will be available at launch:

• GPT-5.6 Terra, designed to balance capability and efficiency in day-to-day work.
• GPT-5.6 Luna, designed to respond faster on straightforward tasks.
• GPT-5.6 Sol, focused on complex scientific and mathematical problems.

OpenAI says GPT-5.6 Sol reaches 83% on FrontierMath Tier 4, a research-level mathematical reasoning benchmark, compared with 72.5% for GPT-5.5. On GeneBench Pro, which evaluates biological data analysis and scientific reasoning, GPT-5.6 Sol Pro solves 31.5% of tasks.

Program accounts will have enterprise-grade privacy and security protections. By default, researchers' data will not be used to train the models.

What they will be able to use it for

The goal is not for OpenAI to choose which scientific problems should be solved, but to put advanced tools in the hands of people who already understand these fields. In practice, a team can use ChatGPT to explore an idea, generate hypotheses, summarize papers or prepare a funding application.

Codex can help write and debug code, analyze datasets and create reproducible workflows. ChatGPT Work is designed for longer projects, such as reviewing scientific literature, drafting papers, finding funding opportunities or preparing materials to communicate results.

There will also be more than 75 specialized tools for life sciences, with features related to genetics, genomics, sequencing, single-cell analysis, protein modeling and drug discovery. Connectors will support work with scientific literature, genomic and clinical databases, satellite imagery, spreadsheets and reference managers.

AI is already entering laboratories

OpenAI says around 1.3 million people use ChatGPT each week for advanced science and mathematics tasks, generating roughly 8.4 million messages. In mathematics, the company says usage has shifted from solving isolated problems to becoming a regular part of some research projects.

There are concrete examples. Physicist Rogerio Jorge and his team use AI to develop open research software for fusion, which is used by companies and national laboratories. In theoretical computer science, Barna Saha, Yinzhan Xu and Christopher Ye used GPT-5.5 Pro to develop a proof about the efficiency limits of computers on high-dimensional geometry problems. The researchers then validated and refined the result themselves.

According to data shared by OpenAI, researchers who use AI the most also tend to delegate longer tasks. Among the top 20% of users in their respective fields, nearly 7% of requests involve work estimated to take four hours or more, compared with 3.5% for the rest.

Who can apply

The initial program is open to researchers at selected academic institutions. They must belong to recognized universities with degree-granting programs and a high level of research activity. They will also need to verify their affiliation, explain what research they are working on and detail how they intend to use AI.

Institutions that already use ChatGPT Edu will coordinate free access through their workspace. OpenAI will also provide training, practical assistance and specialist support to help both those starting from scratch and teams already developing advanced applications.

The initiative is part of a commitment of more than $250 million through 2027 to support external scientific research. It includes NextGenAI, a $50 million program for research institutions, and collaboration with the United States Department of Energy on the Genesis Mission.

What matters is not only that more scientists will have access to a chatbot. The program will make it possible to observe which tasks AI can handle reliably, which still require strict human review and where errors appear. The next phase will depend on that evidence: more researchers may be able to work with these systems, but it will also become clearer when you should trust them and when you should not.