OpenAI opened applications on October 15 for a sponsored GPT-Rosalind plan aimed at academic labs at U.S. universities. A principal investigator or lab lead can apply for up to 30 ChatGPT Business Standard seats at $0 per user per month for up to 12 months. Access to the life sciences model requires a separate trusted-access approval and activation process.

For lab leaders, the useful distinction is between being eligible to apply and having an active workspace in which named members can use GPT-Rosalind. OpenAI's application guide sets out the steps and the limits, including a payment-card requirement during offer setup and no API access under this plan.

The big change

  • What changed: OpenAI has opened a sponsored route for eligible U.S. university labs to seek GPT-Rosalind access through a managed ChatGPT Business workspace. The accepted offer can cover up to 30 Standard seats for up to 12 months.
  • Why it matters: A lab can apply through its PI or designated lead without a university Enterprise contract, provided its institution permits an independent workspace. The application, trusted-access decision and individual member activation each have separate requirements.
  • What to watch: OpenAI has not promised approval to every eligible applicant. Lab leads need to check the accepted offer's end date and charges, their university's data rules and the account route that applies to them.

Which route applies to a lab

The academic offer is currently limited to labs at U.S. universities. A PI or designated lab lead applies on behalf of the lab; members do not submit separate applications. OpenAI says an institutional email address or an existing subscription does not guarantee approval. The lead must supply institutional, lab and intended-use information and identify a trained institutional biosafety contact. OpenAI reviews the lab's affiliation, research activities and the lead's authority to oversee access.

The university's existing OpenAI arrangement determines the application path. A lab without an Enterprise agreement can use the academic-lab application after checking that its university permits an independent ChatGPT Business workspace. A lab at a university with an OpenAI Enterprise agreement should request GPT-Rosalind through that agreement and its university administrator or OpenAI account team. Its existing agreement governs its data terms. Researchers uncertain of their route should check with their institution before setting up a workspace.

Verification, approval and activation are separate

OpenAI's guide instructs the lead to sign in with the ChatGPT account tied to an institutional .edu address, verify the email, provide lab and research information, complete identity verification and attestations, and submit the application. Persona verifies the university and the lead's identity. Completing that verification does not approve the application; OpenAI reviews eligibility and responsible use and may request more information. Applicants can check the status page and should avoid duplicate applications while a review is pending.

After an approval email, the lead uses the status page for setup instructions. Before accepting, OpenAI tells the lead to review the number of covered seats, the end of the complimentary period and any charges. A payment card is required during setup. The lead then completes the linked workspace setup, accepts the access terms, adds eligible members and enables GPT-Rosalind for each one. Joining the workspace alone does not enable the model.

The verified lead is the sole workspace owner and manages permissions. The guide says members must have verified institutional email addresses in the same domain as the approved lead. It directs labs needing more than 30 seats to OpenAI Support; creating a separate paid workspace does not extend the academic offer or transfer GPT-Rosalind approval.

Tools and data boundaries

The Standard seats include ChatGPT and Codex subject to their usage limits, while GPT-Rosalind access follows trusted-access approval and activation. OpenAI says commercially available frontier models are included within plan usage limits. The offer provides no API credits or API usage and does not grant GPT-Rosalind access through the API. Labs seeking to integrate the model into their own software cannot do so through this academic offer.

OpenAI says research data shared through the plan is not used to train its models by default. Its Help Center also says the managed Business workspace has standard data retention and that Zero Data Retention is unavailable under this offer. Depending on controls and law, an administrator may be able to access, export, retain or delete workspace data. Labs should check their own institutional data-handling rules before adding research material. The PI or lab lead remains responsible for membership and approved use; OpenAI says safety monitoring and enforcement continue after activation.

OpenAI's materials describe access conditions and intended scientific uses. BIG CHANGE did not apply, set up a workspace, use GPT-Rosalind or test a scientific result. The practical decision for a lab is whether its institution permits this workspace, who can act as lead and safety contact, and whether the accepted offer and data terms fit its research.

Sources & further reading