OpenAI rolled out GPT‑Rosalind 5.5, a life‑science model that fuses agentic coding with deeper medicinal‑chemistry and genomics expertise. The update delivers sharper drug‑discovery reasoning and broader experimental workflow support. Researchers can now query complex chemistry queries and wet‑lab troubleshooting with a single, unified interface for pharmaceutical teams across institute global collaborations and clinical development projects.

Benchmarking on LifeSciBench shows GPT‑Rosalind outperforms GPT‑5.5 by 27.5 percent on medicinal‑chemistry tasks while cutting token usage by 7.2 percent. GeneBench and LabWorkBench evaluations further reveal 31 percent fewer tokens and 21.6% higher accuracy in genomics analysis, and a 63.2% win rate over GPT‑5.5 in wet‑lab protocol assistance for researchers.

OpenAI also released Life Sciences Research and NGS Analysis plugins that embed the model’s intelligence into repeatable workflows. These tools pull evidence from papers, perform bioinformatics execution, and render sequence, alignment, and structure viewers directly inside Codex, letting scientists trace decisions and validate outputs without leaving the workspace for clinical investigators and data scientists in.

Access remains restricted to vetted organizations through trusted‑access deployment. Early adopters include Novo Nordisk, which will use GPT‑Rosalind to accelerate therapeutic discovery. The rollout marks a significant step toward integrating AI into the full life‑science pipeline, from hypothesis to laboratory execution for patients worldwide and to drive innovation in drug development efforts and clinical trial.