GPT-Rosalind brings GPT-5.5 tool use into life-science workflows
Original: GPT-Rosalind adds GPT-5.5 coding and tool use for life sciences workflows View original →
Life-science AI is moving beyond summarizing papers toward executable workflows. OpenAI posted on June 3, 2026 that GPT-Rosalind is “purpose-built for life sciences research at enterprise scale.” The update combines GPT-5.5-style coding and tool use with stronger intelligence for drug discovery, analysis, design, and experimental work.
“drug discovery, analysis, design, and experimental workflows”
OpenAI’s main account is the primary channel for model and product updates, and the linked article gives benchmark details. On MedChemBench, GPT-Rosalind scored 27.5% versus GPT-5.5 at 25.1% while using 7.2% fewer tokens. On GeneBench, it reached 21.6% versus 20.4% while using 31% fewer tokens. On LabWorkBench, which tests assistance with real wet lab protocols, it scored 63.2% versus GPT-5.5 at 55.8%.
The benchmark framing matters because OpenAI is not presenting a single biology quiz. LifeSciBench draws tasks from six workflow areas: evidence handling, analysis, design and optimization, scientific reasoning, validation and operations, and translation and communication. OpenAI also added Life Sciences Research and Life Sciences NGS Analysis plugins so researchers can connect literature, omics analysis, and interactive sequence, alignment, and structure viewers inside Codex.
What to watch next is access and governance. GPT-Rosalind is available as a research preview for eligible organizations through trusted access, and OpenAI named Novo Nordisk as a partner exploring frontier AI for medical research. In this domain, provenance, expert review, and biosecurity controls will matter as much as raw scores. Source: OpenAI on X · GPT-Rosalind article
Related Articles
OpenAI put GPT-Rosalind into research preview for qualified life-science teams, pairing a domain model with a Codex plugin that connects to more than 50 tools and data sources. The strongest signal is not the branding: OpenAI says best-of-ten submissions ranked above the 95th percentile of human experts on one Dyno Therapeutics RNA prediction task.
OpenAI is moving model specialization into scientific work rather than generic chat. GPT-Rosalind is framed for protein reasoning, chemical reasoning, genomics, biochemistry and tool use, with access starting as a research preview for qualified customers including Amgen and Moderna.
NVIDIA says Lilly has brought LillyPod online, a pharma-owned AI factory built with 1,016 Blackwell Ultra GPUs and more than 9,000 petaflops of AI performance. The platform is intended to accelerate model training and simulation across genomics, small molecules, and clinical-development workflows.