OpenAI offers US university labs 30 free ChatGPT seats and GPT-Rosalind access

What OpenAI announced on 15 October

OpenAI announced the Rosalind Academic Plan on 15 October 2026 under the headline 'Bringing GPT-Rosalind to academic research labs.' The plan offers eligible labs sponsored access to GPT-Rosalind, which OpenAI calls its dedicated life-sciences model, along with the tools it pairs with that model for discovery work.

OpenAI frames the offer around the role university research has played in earlier breakthroughs. The post cites the Human Genome Project, the curiosity-driven work on bacterial immune systems that led to CRISPR-Cas9, and decades of university mRNA research that preceded COVID-19 vaccines. Its pitch is that lowering the cost of access gives more researchers room to pursue ideas whose benefits could reach beyond their own labs.

The mechanics are spelled out in a Rosalind Academic Plan help-centre article, which was updated on launch day and covers eligibility, the application, seat management and data handling in more detail than the announcement. Unite.AI's write-up the same evening matches those terms.

What an approved lab gets

OpenAI lists three things for eligible labs: up to 30 Business Standard seats at $0 per user per month for up to 12 months; access to all commercially available frontier OpenAI models, subject to plan usage limits; and access to GPT-Rosalind following trusted-access approval. The help article adds that a Standard seat includes ChatGPT and Codex in a shared workspace for reviewing literature, working with research data and developing analysis code.

The fine print matters. OpenAI says a payment card is required during setup even though the seats carry no charge for the 12 months. The complimentary offer does not remove ChatGPT, Codex or GPT-Rosalind usage limits. It includes no API credits or API usage, and approval does not grant GPT-Rosalind API access. The help article tells labs to check the accepted offer's end date and terms for what happens after the free period.

On training data, the announcement says research data shared through the Rosalind Academic Plan is not used to train OpenAI's models by default. The help article separately notes that the plan is distinct from OpenAI's existing ChatGPT for Academic Researchers programme.

Who can apply and what the application asks for

The plan is currently open only to academic labs at US universities. Principal investigators and lab leads at universities without an OpenAI Enterprise agreement apply through OpenAI's trusted-access process. Labs at universities that already have an Enterprise agreement must request GPT-Rosalind through their university's Enterprise access process instead. Unaffiliated researchers and commercial organisations are not eligible, and only US accounts qualify.

Applicants sign in with a ChatGPT account on an institutional .edu email, verify it, and submit the institution, department, lab name and website; the PI's profile, role and research focus; the intended use of GPT-Rosalind; the people who need access, with their emails and affiliations; any existing workspace, access locations and data-handling requirements; and the name, title and contact details of a trained biosafety contact at the institution.

The help article points applicants to their Environmental Health and Safety, Research Safety or biosecurity office to find that contact, who must be an institution-designated official such as a biosafety officer or a Responsible Official in a select-agent programme. A PI can fill the role if the institution designates them. OpenAI uses Persona to verify the university and the PI's identity, and warns that passing Persona, holding a university email or having an existing subscription does not guarantee approval. Applicants are also told to check whether their university permits an independent ChatGPT Business workspace.

How trusted access and ongoing checks work

Before activating access, OpenAI says it verifies the institution's legitimacy and the PI's identity, reviews the lead's professional background, appointment and authority to oversee lab access, assesses the lab's affiliation, research activities and intended use, and sets responsibilities for both the lab lead and the designated institutional safety contact.

Safeguards continue after activation through monitoring and enforcement. For registered high-sensitivity use cases, OpenAI says its safety checks take the approved research context into account, and that it may follow up with researchers to clarify the work and confirm authorisation. According to the help article, access can be revoked if the safety monitoring and responsible-use requirements are no longer met.

The announcement also publishes an access-comparison example. On a prompt requesting a 20-variant sequence and presentation matrix for a second-generation Nipah antiviral lipopeptide campaign, plus modelling targets and paired assays, OpenAI says GPT-Rosalind produced a candidate matrix, a modelling plan and paired study plans with controls, while a model it labels Mainline Astra withheld the sequence-level designs and gave only high-level context. OpenAI notes the published assessments address whether the text fulfilled the requested deliverables and do not verify scientific correctness, feasibility or whether a method actually works.

Workspace and data rules

Participating labs operate in a managed ChatGPT Business workspace. The help article says standard data retention applies by default and that Zero Data Retention is not available under this offer, so a workspace admin can access, export, retain or delete content under the normal Business rules.

The verified PI or lab lead is the sole workspace owner and cannot appoint additional owners or admins. Joining the workspace does not by itself grant GPT-Rosalind; the PI enables it per member in admin settings. Thirty seats is the maximum in a single workspace, every member must use a verified institutional email on the same domain as the approved PI, and a separate paid workspace does not extend the offer or transfer approval.

What GPT-Rosalind and Rosalind Workbench are

OpenAI describes GPT-Rosalind as its dedicated life-sciences model, combining frontier reasoning with a Rosalind Workbench that orchestrates specialised tools across medicinal chemistry, genomics, wet-lab assistance and other scientific work. Its product page describes the Workbench as an orchestrated workspace in research preview, with an Explore mode for general scientific questions and a Research mode for more complex biological questions. The model can also be used inside Codex.

The listed research areas are drug and protein design; disease detection and monitoring; vaccines and immune responses; genetic and RNA therapies; agriculture and animal health; ecology and conservation; and antimicrobial resistance. OpenAI first introduced GPT-Rosalind in a product post on 7 October 2026; the 15 October plan is about who can access it and how, not a new model.

The academic offer lands amid a wider push of AI into the lab bench. We have covered Google's $36 million AI-for-science commitment across 17 organisations and DeepMind's AlphaProtein enzyme work; OpenAI's plan tackles the access and governance side rather than the model itself.

Early academic users OpenAI names

OpenAI names several labs already using GPT-Rosalind, on work it describes as ranging from protein design to understanding neurodegeneration. At UCSF, it says the Zhao Lab is using the model for quantitative drug development. OpenAI quotes the lab saying GPT-Rosalind helped it gather drug physicochemical parameters faster and that PubChem and NCBI tools helped it work across ten drugs while refining the distinction between predicted and experimentally measured values.

Dr Mathieu Bourdenx of the Bourdenx Lab is quoted saying he uses GPT models throughout his research to iterate on ideas faster. Dr Brian Hie of Stanford University and the Arc Institute is quoted saying Rosalind will help improve current discovery and design tasks and apply them at larger scales than previously possible while still providing useful insights to biological researchers. These are testimonials OpenAI selected and published, not independent evaluations.

Limits and open questions

The headline is generous free access, but the gates are the story. GPT-Rosalind itself is behind trusted-access approval, limited to US university labs, tied to a biosafety contact, and excluded from the API. A payment card is required even for the free seats, and OpenAI has not published approval timelines, how many labs it expects to admit, or what happens to pricing after the 12-month window.

The safety framing is also OpenAI's own. The Nipah comparison and the high-sensitivity monitoring are described by OpenAI without independent review, and the company itself cautions that its published assessments do not verify whether a method works. For labs weighing the offer, the trade is real frontier-model access at no seat cost against a managed workspace without Zero Data Retention and a GPT-Rosalind gate they do not control.