Key Highlights
- On April 17, 2026, OpenAI introduced GPT-Rosalind, an artificial intelligence model specifically engineered for biological research and pharmaceutical development
- The technology targets reduction of the traditional 10–15 year drug approval cycle in the United States through enhanced early-phase research capabilities
- Qualified researchers can access the model through ChatGPT, Codex, and API platforms, complemented by a complimentary plugin integrating over 50 scientific databases
- Collaborative partnerships span major industry players: Amgen, Moderna, Thermo Fisher Scientific, Novo Nordisk, and the Allen Institute
- Thermo Fisher Scientific maintains a GF Score of 86/100, while company insiders divested $8.2 million in equity over the previous quarter
On April 17, 2026, OpenAI revealed GPT-Rosalind, an artificial intelligence system engineered specifically for advancing biological sciences, pharmaceutical research, and translational medical applications.
The naming honors Rosalind Franklin, whose pioneering crystallography research proved instrumental in uncovering DNA’s double helix structure.
The standard pharmaceutical approval pathway in the United States spans a decade to a decade and a half. According to OpenAI, GPT-Rosalind targets acceleration of the initial research phases within this extended timeline.
The artificial intelligence system helps scientific teams navigate vast collections of published research, biological databases, and experimental datasets with enhanced efficiency.
Capabilities include systematic literature analysis, scientific hypothesis formulation, experimental protocol development, and biological sequence functional prediction.
GPT-Rosalind currently operates as a research preview accessible via ChatGPT, Codex, and OpenAI’s application programming interface. Present access remains restricted to vetted enterprise partners within United States borders.
OpenAI simultaneously launched a complimentary Life Sciences research extension for Codex, enabling integration with more than 50 specialized scientific databases and analytical platforms.
Performance Metrics and Benchmarking
BixBench testing, which evaluates practical bioinformatics applications, positioned GPT-Rosalind at 0.751, surpassing both GPT-5.4’s score of 0.732 and Grok 4.2’s result of 0.698.
LABBench2 evaluations showed GPT-Rosalind exceeding GPT-5.4 performance across 6 of 11 specialized research applications. The most significant advancement appeared in DNA cloning protocol generation.
Additional validation involved Dyno Therapeutics using proprietary RNA sequence data. Results placed the model above the 95th percentile when compared against human expert predictions.
Strategic Industry Collaborations
OpenAI has established partnerships with Amgen, Moderna, Novo Nordisk, Thermo Fisher Scientific, NVIDIA, Oracle Health and Life Sciences, the Allen Institute, Benchling, and UCSF School of Pharmacy.
Amgen’s Senior Vice President overseeing AI and Data initiatives highlighted that partnering with OpenAI enables application of sophisticated technologies through novel approaches, ultimately accelerating therapeutic delivery to patients.
Thermo Fisher Scientific, featured among the collaborative partners, commands a market capitalization approximating $191.76 billion. The corporation specializes in scientific instrumentation, laboratory equipment manufacturing, and life science reagent production.
Thermo Fisher’s GF Score registers at 86 out of 100, supported by a Profitability Rank of 8/10 alongside a Growth Rank of 7/10.
Recent insider trading activity at Thermo Fisher reveals $8.2 million in share sales during the preceding three-month period, with zero insider purchase transactions recorded during the same timeframe.
Microsoft provides backing for the model, continuing its established investment partnership with OpenAI.



