OpenAI unveiled Rosalind Biodefense, a new initiative providing trusted developers access to GPT-Rosalind for building defensive biosecurity tools. The program targets pandemic preparedness applications including epidemiological modeling, early detection systems, and screening infrastructure for DNA synthesis. OpenAI will sponsor access and provide launch support for qualified teams working on public health interventions.

The company is also expanding GPT-Rosalind access to select U.S. government and allied partners with approved public health missions. This builds on their July 2025 release of ChatGPT agent, the first model designated as High Capability in biology under their Preparedness Framework. Since then, they have refined safeguards and conducted pre-deployment evaluations with external testing groups.

Fourth Eon Biosecurity is among the initial organizations using GPT-Rosalind to develop AI-native screening systems that analyze DNA sequences for potential threats. These partnerships demonstrate how advanced AI can strengthen prevention workflows and support researchers, infrastructure operators, and communities in preparing for biological risks.

The initiative reflects OpenAI's layered resilience approach: investing in preparedness evaluations, bio-specific capability assessments, and safer model behavior for dual-use requests. By expanding access through trusted models, they aim to accelerate defensive research while maintaining appropriate safety controls for advanced biological capabilities.