OpenAI announced the debut of GPT-Rosalind on April 16, 2026, a specialized artificial intelligence model engineered specifically for the life sciences sector. Named after the British chemist Rosalind Franklin, the model is designed to act as a dedicated intelligence layer for biochemistry, drug discovery, and translational medicine. According to the company, GPT-Rosalind is being released as a limited-access research preview for qualified Enterprise customers in the United States, marking a significant move toward domain-specific vertical integration for the AI firm.
The model is optimized for scientific workflows, integrating capabilities across protein engineering, genomics, and molecular chemistry. In internal evaluations, GPT-Rosalind achieved a 0.751 pass rate on BixBench, a benchmark for bioinformatics and data analysis. It also outperformed OpenAI’s general-purpose GPT-5.4 model on LABBench2, which measures proficiency in research tasks such as literature retrieval and experimental protocol design. OpenAI stated that the model is intended to help researchers synthesize evidence and generate biological hypotheses, potentially shortening the traditional 10-to-15-year drug development cycle.
OpenAI confirmed it is collaborating with several major industry partners to deploy the model, including Amgen Inc., Moderna Inc., and Thermo Fisher Scientific. These organizations are using GPT-Rosalind to analyze large datasets and convert scientific findings into clinical applications. Alongside the model, OpenAI introduced a Life Sciences research plugin for Codex, which connects the intelligence layer to more than 50 external scientific tools and data sources. This allows researchers to query databases and plan experiments directly within the AI environment.
Brad Lightcap, OpenAI’s Chief Operating Officer, noted during a press briefing that the model represents a shift from general assistance to domain-specific reasoning. Lightcap emphasized that by providing a tool optimized for the complexities of biology, the company aims to reduce the massive costs and timeframes associated with early-stage discovery. To support these sensitive workflows, the model is built on infrastructure that meets HIPAA and SOC 2 Type II compliance standards. It also features a specialized Verification Layer designed to cross-reference AI-generated findings with peer-reviewed literature to ensure scientific accuracy.
The launch places OpenAI in direct competition with Anthropic, which released its own Claude for Life Sciences model approximately five months ago. While GPT-Rosalind is currently restricted to gated access, OpenAI indicated that it plans to expand availability as it refines the model’s safety protocols. The company did not disclose specific pricing for the new model but confirmed it will be offered as a premium service within its Enterprise suite.