OpenAI Announces GPT-Rosalind, a Frontier Reasoning Model for Life Sciences

OpenAI has announced GPT-Rosalind, their first frontier reasoning model specialized for life sciences. The model aims to compress the drug discovery process by handling tasks like evidence synthesis, hypothesis generation, and experimental planning.

Details

GPT-Rosalind is designed to address the challenges in drug discovery, which typically takes 10-15 years from target identification to regulatory approval. The model aims to streamline this process by handling tasks like evidence synthesis, hypothesis generation, experimental planning, and sequence-to-function interpretation. According to OpenAI, GPT-Rosalind has achieved the highest publicly reported score on the BixBench bioinformatics benchmark and outperformed GPT-5.4 on 6 out of 11 tasks in the LABBench2 evaluation. In a private evaluation by Dyno Therapeutics, the model's performance on an RNA sequence-to-function task exceeded the 95th percentile of 57 human experts. However, the model is not publicly available and is only accessible through OpenAI's Trusted Access Program, which has strict requirements around beneficial use, governance, and security. To provide broader access, OpenAI has also released a free Life Sciences Research Plugin for Codex, which offers access to public databases and tools. The launch partner list includes major pharmaceutical companies, research labs, and consulting firms, indicating real-world deployment of the technology.