GPT‑Rosalind for drug discovery
- OpenAI announced GPT‑Rosalind, a specialized reasoning model aimed at drug discovery and genomics workflows. - Early adopters reportedly include Amgen, signaling enterprise interest in vertical reasoning models. - Specialized models like Rosalind suggest vendors are segmenting by domain, creating options for startups that need domain-specific reasoning (thesauditimes.net).
Drug discovery starts with a basic problem: scientists have to connect genes, proteins, chemicals, papers, and lab data before they can decide what to test. OpenAI said on April 16 it built GPT‑Rosalind to help with that work in biology and medicine. (openai.com) OpenAI described GPT‑Rosalind as a “frontier reasoning model” for drug discovery, genomics analysis, protein reasoning, and translational medicine. The company said the model is available in research preview through ChatGPT, Codex, and the application programming interface for qualified customers. (openai.com) Genomics is the study of DNA, and proteomics is the study of proteins that do the cell’s work. OpenAI said Rosalind is tuned for multi-step tasks such as target discovery, target validation, pathway analysis, literature synthesis, experimental planning, and data analysis. (openai.com; help.openai.com) OpenAI put the timing in drug-development terms: in the United States, it said, moving from target discovery to regulatory approval for a new drug usually takes about 10 to 15 years. The company said faster work at the front end can improve target selection, biological hypotheses, and experiment quality later on. (openai.com) The model is not a general public launch. OpenAI said GPT‑Rosalind is being released through a “trusted access” program for eligible institutions, with screening for beneficial use, governance, misuse prevention, access controls, and enterprise security. (openai.com) OpenAI also tied the model to software around the model. The company said it is releasing a Life Sciences research plugin for Codex that connects to more than 50 scientific tools and data sources, so researchers can move between papers, databases, and internal systems in one workflow. (openai.com) Amgen is one of the companies OpenAI named as a customer working with GPT‑Rosalind, alongside Moderna, the Allen Institute, and Thermo Fisher Scientific. That fits Amgen’s existing research strategy: the company says it already uses artificial intelligence across research and development, including work on omics data and clinical development. (openai.com; amgen.com) Omics is a catch-all term for large biological datasets from genes, RNA, and proteins. Amgen said in April 2025 that it uses genomics, transcriptomics, and proteomics to prioritize drug targets and design more targeted medicines, which is the kind of workflow Rosalind is meant to support. (amgen.com) OpenAI named the model after Rosalind Franklin, whose research helped reveal DNA’s structure. The immediate test is narrower than the branding: whether a specialized model can help lab teams make faster, better-grounded decisions before they spend money on experiments. (openai.com)
