OpenAI opens frontier AI program to 100,000 academic researchers

OpenAI sets a 100,000-researcher target

OpenAI has introduced ChatGPT for Academic Researchers, giving 100,000 scientists, mathematicians and engineers at selected institutions free access to frontier models. It starts with 10,000 researchers this summer and is planned to expand through 2027.

Access is already available at the Institute for Advanced Study and École normale supérieure. Approved researchers may invite up to four institutional collaborators, each verified and counted toward the 100,000 accounts.

Models, tools and safeguards

Participants receive ChatGPT, ChatGPT Work and Codex, including the GPT‑5.6 family at launch, plus expanded deep research, higher limits and larger context windows.

OpenAI positions GPT‑5.6 Terra for efficient everyday research, Luna for lighter tasks and Sol for difficult scientific and mathematical work. Sol scored 83% on FrontierMath Tier 4, versus 72.5% for GPT‑5.5; Sol Pro solved 31.5% of GeneBench Pro tasks.

More than 75 life-science skills cover fields from genomics and single-cell analysis to protein modelling and drug discovery. Connectors link to scientific literature, databases, notebooks and reference managers.

Workspaces have business-grade privacy and security protections, and data is not used to train OpenAI's models by default. Codex supports coding and reproducible analysis; ChatGPT Work supports longer projects such as grant applications.

Evidence of research use

OpenAI says about 1.3 million people use ChatGPT for advanced science and mathematics weekly, generating 8.4 million messages. From June 21 to July 20, 2026, the top 20% of users in a field assigned nearly 7% of requests to tasks estimated at four hours or more, versus 3.5% among peers.

The program includes tiered training, specialist support and opportunities to exchange methods. At institutions using ChatGPT Edu, free access will be coordinated through the existing workspace.

Eligibility and business implications

Applicants must be researchers at selected, recognized degree-granting institutions with high research activity. They must verify their affiliation and explain their active research and intended scientific use.

For research-led organisations, the practical implication is to authenticate users, define suitable workflows, review outputs and confirm privacy defaults before widening access to frontier AI tools.