The model has been developed to support researchers working across disciplines such as biochemistry, genomics and protein engineering, where increasing volumes of data and fragmented workflows continue to slow progress.
According to the company, it currently takes between ten and fifteen years for a new drug to progress from target discovery to regulatory approval in the US, with only around one in ten candidates entering clinical trials ultimately reaching the market.
GPT-Rosalind is designed to address bottlenecks at the earliest stages of discovery by assisting with evidence synthesis, hypothesis generation, experimental planning and data analysis.
By improving decision-making early in the pipeline, the company believes downstream success rates and development timelines could be significantly improved.
Named after Rosalind Franklin, OpenAI stated that the model is optimised for multi-step scientific workflows, enabling researchers to analyse molecular interactions, interpret genomic data and design experiments more efficiently.
It also integrates with a new Life Sciences research plugin that connects users to more than 50 scientific databases and tools spanning functional genomics, protein structure and clinical evidence.
The model is being launched as a research preview through a "trusted access programme," with availability initially limited to qualified enterprise users in the US.
Early collaborators include major pharmaceutical and life sciences organisations such as Amgen, Moderna and Thermo Fisher Scientific.
OpenAI also emphasised in its statement that GPT-Rosalind is intended to augment human expertise, rather than replace it.
While the model can support complex reasoning and identify patterns across datasets, researchers will remain responsible for validating findings and ensuring experimental accuracy.
The launch comes amid growing industry interest in applying artificial intelligence to drug discovery and translational medicine, where digital tools are increasingly used to identify targets and optimise compounds.
However, relatively few AI-designed drugs have progressed to late-stage clinical trials, highlighting the challenges of translating computational insights into real-world therapies.
Alongside its potential benefits, the deployment of advanced AI models in biology has raised concerns around misuse, particularly in relation to the design of harmful biological agents.
In response, OpenAI stated that GPT-Rosalind had been developed with enterprise-grade security controls, strict access management and governance requirements to ensure responsible use in regulated environments.
Looking ahead, the company plans to expand the model's capabilities and continue working with research institutions to evaluate its real-world impact.
OpenAI believes systems such as GPT-Rosalind could become increasingly valuable tools in helping scientists move more efficiently from data to discovery—and ultimately to new treatments for patients.