OpenAI has launched a new initiative, ChatGPT for Academic Researchers, which will give scientists, mathematicians and engineers at selected academic institutions free access to its AI models. The scheme forms part of a wider pledge by the company to commit more than $250 million (£195 million) through 2027 to support external scientific research, and reflects growing efforts by AI firms to embed their tools within research institutions ahead of any future commercial opportunities this may create.

The programme will begin with 10,000 participants this summer, expanding to 100,000 by 2027, with researchers gaining hands-on support from OpenAI, access to its latest GPT-5.6 Sol Pro model, and the ability to invite four collaborators each from their institution. OpenAI says researchers' data will not be used to train its models by default. The move follows a similar pattern to ChatGPT Edu and January's launch of Prism, a tool for verifying citations and formatting in academic documents, underlining OpenAI's broader push to make itself indispensable across education and research.