OpenAI said Thursday it is launching a new program called Rosalind Biodefense, aimed at giving trusted developers access to its biology-focused AI tools for pandemic preparedness and other defensive life sciences applications. The company is also widening access to GPT-Rosalind for select U.S. government and allied partners working on public health and biodefense missions.
The announcement marks OpenAI’s latest effort to position advanced AI as a tool for improving biological resilience, while maintaining tight controls around who can use the models. The company said the new program is designed to help vetted organizations build applications that support preparedness before a biological threat emerges.
OpenAI said Rosalind Biodefense will sponsor access to GPT-Rosalind, its frontier reasoning model for life sciences research, for organizations developing defensive tools. The company described potential use cases including epidemiological modeling, early detection systems, screening, preparedness planning, non-pharmaceutical interventions and other public-health-related capabilities.
At launch, OpenAI said it is backing a first group of organizations working across different parts of the biodefense stack. One of them, Fourth Eon Biosecurity, is building screening infrastructure for DNA synthesis orders, with the goal of helping labs and companies spot unsafe or malicious requests, including novel designs that might evade traditional checks.
OpenAI framed the effort as part of a broader push to ensure that frontier AI helps defenders as much as it helps researchers. The company said the program is intended to support projects that improve the speed, quality or scale of defensive work, including literature synthesis, protocol design support, model building, data harmonization, simulation, decision support and scientific communication.
The company said applications are open to academic, nonprofit, government-affiliated and mission-driven commercial groups, as long as the work has clear public benefit. It also said the program is open to qualified applicants globally.
Alongside the developer program, OpenAI is extending trusted access to GPT-Rosalind for select public institutions. The company said approved U.S. government and allied partners will be able to use the model for high-impact defensive work such as early warning systems, outbreak response planning, diagnostics, preparedness and medical countermeasure development.
Among the partners named in the announcement is Lawrence Livermore National Laboratory, which is using AI to support biopreparedness and bioresilience efforts. OpenAI said the lab is combining AI with supercomputing, simulation and laboratory testing to help evaluate possible medical countermeasures for emerging biological threats.
OpenAI also said Johns Hopkins Applied Physics Laboratory plans to integrate GPT-Rosalind into a protein-engineering platform to speed screening of mutant enzymes for therapeutics, countermeasure development and the characterization of emerging biothreats. In addition, the company is extending access to the Coalition for Epidemic Preparedness Innovations, or CEPI, which is focused on its 100 Days Mission to accelerate vaccine development against epidemic and pandemic threats.
The company said the new access model is designed to keep safety, security and accountability controls in place while still enabling qualified institutions to use the technology for defensive purposes. OpenAI added that it has been building safeguards for biology-related capabilities through capability assessments, red teaming, monitoring, security controls and other measures.
OpenAI said the Rosalind Biodefense Program is an initial step in a broader strategy and that it expects to expand how trusted partners use GPT-Rosalind over time. The company said it will continue refining access pathways and safeguards as it learns from early deployments and works with partners in the U.S. and abroad.