OpenAI Gives 100,000 Researchers Free Access to GPT-5.6

OpenAI announced on July 29, 2026 a new program called ChatGPT for Academic Researchers, aimed at giving free access to its most powerful language models to 100,000 researchers worldwide. It's one of the company's largest efforts yet aimed directly at the academic community, rather than business or government.

Who can apply

Access isn't open to everyone. OpenAI has limited the program to faculty members and postdoctoral researchers employed at recognized, degree-granting institutions with high research activity. Applicants must confirm their university affiliation through the SheerID verification system and show a publication from the last three years posted on arXiv, bioRxiv, or ChemRxiv.

Qualifying fields include biological sciences, chemistry and materials science, computer science, earth and planetary sciences, engineering, mathematics, and physics. It's a fairly broad but still selective set of criteria, meant to filter out applications unrelated to active research work.

What researchers get

Qualifying researchers get access to the entire GPT-5.6 model family, including GPT-5.6 Sol Pro, OpenAI's current flagship model built for complex queries and long-running tasks. The package also includes Codex, an agentic coding environment, and ChatGPT Work for automating task sequences.

The program also includes an expanded deep research feature, launched in February 2026, which can search hundreds of web pages or focus solely on academic journals. Researchers also get higher usage limits and larger context windows than standard ChatGPT, and OpenAI is providing training on using the tool for research work as well as technical help integrating it with existing scientific software.

Data submitted through the program is not used to train OpenAI's models, and workspaces are meant to carry the same privacy protections as those for enterprise customers. Despite the broad access to the models themselves, the underlying neural network weights remain closed, so the program doesn't change anything about OpenAI's openness as such.

First wave and timeline

The first researchers will start using the program later this summer, with pilot institutions including the Institute for Advanced Study and the Ecole normale superieure. OpenAI says the full 100,000 slots will be rolled out gradually, with a target completion date of 2027.

Context of investment in science

The program fits into OpenAI's broader strategy toward academia. The company entered higher education back in 2024 with ChatGPT Edu, which included safety controls and usage analytics dashboards for universities. In May 2026, OpenAI gave universities $50 million worth of computing resources, and also supported research projects tied to the U.S. Department of Energy.

The stated commitment of more than $250 million for external scientific projects through 2027 is the most concrete figure OpenAI has given so far regarding its involvement in science. By comparison, the sums the company spends on computing infrastructure run into the tens of billions of dollars, making the researcher program a relatively small, but clearly image-building, part of its strategy.

What this means for researchers

For labs and departments without a budget for commercial AI licenses, free access to OpenAI's most powerful model could meaningfully speed up work on publications, data analysis, or literature reviews. At the same time, the program deepens academia's dependence on the infrastructure of a single private company, something some researchers may view with caution, especially since access to the model weights themselves remains closed.

Polish researchers at research-intensive institutions can in theory apply under the same terms as everyone else, though the first wave of the program is focused on universities such as the Institute for Advanced Study and the Ecole normale superieure. Broader availability, reaching the full 100,000 slots, is expected only in later stages through 2027.