OpenAI has launched GPT-Rosalind, a frontier reasoning model specifically designed to accelerate research in biology, drug discovery, and translational medicine. Named after Rosalind Franklin, the model addresses the lengthy and complex nature of the drug development pipeline, which currently takes 10 to 15 years from target discovery to regulatory approval. By optimizing workflows across chemistry, protein engineering, and genomics, GPT-Rosalind aims to help scientists generate better hypotheses, synthesize evidence more efficiently, and execute multi-step experiments faster. The model is now available as a research preview for qualified customers through OpenAI's trusted access program on ChatGPT, Codex, and the API. Unlike general-purpose models, GPT-Rosalind excels in tasks requiring deep domain reasoning, including analyzing chemical reaction mechanisms, predicting protein structure and mutation effects, and interpreting DNA sequence phylogenetics. In public benchmarks, the model demonstrated leading performance on BixBench, a bioinformatics and data analysis evaluation, and outperformed previous versions on six out of 11 tasks on LABBench2. Notably, in an evaluation with Dyno Therapeutics, the model ranked above the 95th percentile of human experts in predicting RNA sequence-to-function relationships. To support practical application, OpenAI introduced a freely accessible Life Sciences research plugin for Codex. This tool connects AI models to over 50 public scientific databases, literature sources, and specialized tools, acting as an orchestration layer for complex, multi-step research questions. While eligible enterprise users can leverage the plugin with GPT-Rosalind for advanced biological reasoning, the tool is also compatible with OpenAI's mainline models for broader use. Access to the GPT-Rosalind model is governed by strict safety and security protocols. The launch utilizes a trusted-access deployment structure for U.S. enterprise customers, ensuring that only organizations conducting legitimate scientific research with clear public benefit can participate. Eligibility requires adherence to strong governance, compliance with misuse-prevention controls, and operation within secure, well-managed environments. During the research preview, usage does not consume existing tokens, though pricing details will be announced as the program expands. OpenAI is collaborating with leading industry partners such as Amgen, Moderna, the Allen Institute, and Thermo Fisher Scientific to integrate the model into real-world discovery workflows. Future development will focus on expanding the model's capabilities in long-horizon, tool-heavy scientific tasks, with ongoing partnerships involving national laboratories like Los Alamos. The company views this release as the start of a long-term commitment to building AI that can help scientists transition rapidly from questions to evidence, and ultimately to new treatments for patients.