OpenAI opens free GPT-5.6 and Codex program for 100,000 researchers

The first 10,000 researchers will receive access this summer, but eligibility is restricted to selected research-intensive academic institutions

OpenAI’s new program will provide selected academic researchers with free access to ChatGPT, Codex and frontier AI models

OpenAI has opened applications for a free AI program that will give up to 100,000 academic researchers access to ChatGPT, ChatGPT Work, Codex and its frontier models through 2027.

ChatGPT for Academic Researchers will begin with 10,000 scientists, mathematicians and engineers this summer. Access is already available at the Institute for Advanced Study and École normale supérieure, with the wider program targeting researchers working across scientific disciplines, mathematics and engineering.

The offer forms part of an OpenAI commitment of more than $250 million through 2027 to support external scientific research and discovery.

Free access comes with eligibility restrictions

The program is not open to every researcher or university. Applicants must work at selected, recognized degree-granting colleges or universities with a high level of research activity.

Researchers will need to verify their institutional affiliation and provide details of their active research and intended scientific use.

Approved participants can invite up to four collaborators, but each must work at the same institution, complete the affiliation checks and count toward the program’s 100,000-account allocation.

For universities already using ChatGPT Edu, access awarded through the initiative will be coordinated through the institution’s existing workspace.

“We believe the benefits of frontier AI should not be concentrated in a few companies and well-resourced labs,” OpenAI states. “Researchers know their fields best. Our role is to put powerful tools in their hands and help accelerate their work.”

GPT-5.6 access, larger context windows and research support

Participants will receive access to OpenAI’s GPT-5.6 family, including GPT-5.6 Sol Pro at launch, alongside expanded deep research, higher usage limits and larger context windows.

The company says ChatGPT and Codex can support scientific work ranging from genomic analysis, protein modeling and hypothesis generation to literature reviews, grant applications, coding, data analysis and research publishing.

More than 75 life science skills will be available across areas including genetics, genomics, sequencing, single-cell analysis, protein modeling and drug discovery. Connectors will provide access to scientific literature, public genomic and clinical databases, satellite imagery, computational notebooks, data platforms and reference managers.

Codex is intended to help researchers write and debug code, analyze datasets and create reproducible workflows. ChatGPT Work is positioned for longer projects including literature reviews, funding searches, grant applications and manuscript preparation.

Each workspace will include business-grade privacy and security protections. OpenAI states that participants’ research data will not be used to train its models by default.

The package also includes training for researchers with different levels of AI experience and hands-on assistance from specialists familiar with research workflows. Participants will have opportunities to exchange approaches with other researchers and provide feedback on where the models are useful or fall short.

OpenAI points to rising research use

OpenAI says approximately 1.3 million people now use ChatGPT for advanced science and mathematics each week, generating around 8.4 million messages.

The company also recorded a sharp increase in arXiv mathematics papers acknowledging ChatGPT during 2026. The number rose from 14 in February to 100 between July 1 and July 21, although the July total covers an incomplete month.

Its analysis found that researchers in the top 20 percent of AI use within their field were almost twice as likely as their peers to give AI tasks estimated to require at least four hours of human work. Nearly 7 percent of their requests met that threshold, compared with 3.5 percent among other researchers in the same field.

ChatGPT for Academic Researchers sits alongside OpenAI’s $50 million NextGenAI initiative and its work with the US Department of Energy’s Genesis Mission.

Applications are now open to qualifying researchers at selected academic institutions, with no closing date stated.