OpenAI is building a dedicated ChatGPT subscription plan for scientific institutions, aiming to bring its AI tools deeper into the world of academic and corporate research. Code references discovered in OpenAI’s web application on June 17 point to a product called “ChatGPT for Science,” targeting universities, national laboratories, and corporate R&D teams.

No official pricing or launch date has been announced.

What we know so far

The ChatGPT for Science plan appears designed to serve researchers across biology, chemistry, physics, and materials science. Access is expected to follow a trusted-access model, requiring institutional verification and adherence to service agreements.

The product sits on a trajectory that’s been visible for a while now. OpenAI launched ChatGPT Edu in May 2024, giving universities a tailored version of its AI for academic use. In January 2026, the company rolled out the free Prism workspace for scientists. Then in April 2026, it launched GPT-Rosalind, a model specifically designed for life sciences research.

ChatGPT for Science looks like the next logical move: a comprehensive subscription tier that bundles scientific capabilities into a single offering that institutions can procure through their existing vendor relationships.

OpenAI’s institutional playbook

Universities and national labs have procurement offices, compliance requirements, data governance policies, and budgets that move on academic calendars. The institutional verification requirement, the service agreements, and the emphasis on trusted access are signals that the company is building for procurement committees, not individual researchers.

The competitive landscape here is worth noting. Google DeepMind has been making aggressive moves in scientific AI, particularly with protein structure prediction and materials discovery. Anthropic has been positioning Claude for professional and research use cases. Meta’s open-source Llama models have found homes in academic settings where budget constraints make paid subscriptions a harder sell.

What this means for the research economy

Universities and national labs have traditionally spent their software dollars on tools like MATLAB, specialized databases, and statistical packages. AI subscriptions represent an entirely new budget category.

The absence of a confirmed launch date means institutions should be watching for an official announcement rather than making procurement decisions today. The code references and the clear product trajectory suggest this isn’t a speculative concept — it’s a product in testing, following a playbook OpenAI has already executed with ChatGPT Edu and GPT-Rosalind.