On Thursday, April 16th, OpenAI officially launched an early version of its AI model specifically designed for life science research - GPT-Rosalind. Named after Rosalind Franklin, the discoverer of DNA structure, this model aims to accelerate the drug development process using generative AI. By extracting insights from massive biochemical data, GPT-Rosalind can assist researchers in evidence synthesis, hypothesis generation, experimental planning, and protein engineering analysis, aiming to more efficiently convert long laboratory research into medical applications for patients.

Currently, the model is available as a "research preview" to select enterprise customers. The initial partners include pharmaceutical giant Amgen, vaccine manufacturer Moderna, Thermo Fisher Scientific, and non-profit institution Allen Institute. Technically, GPT-Rosalind has been deeply fine-tuned for genomics and chemistry fields, performing well in scientific benchmarks such as LABBench2, and showing potential to surpass human experts in certain prediction tasks. In addition, OpenAI simultaneously launched a complementary life sciences plugin, supporting links to over 50 professional scientific tools and databases.
This move marks OpenAI's formal entry into the scientific AI track led by Google (DeepMind) and Anthropic. As leading AI laboratories focus on medical data analysis and new drug assistance development, AI has evolved from a general chat tool into a research infrastructure. Although there is still observation regarding "AI replacing mature R&D experience," OpenAI's deep layout has caused a stir in the pharmaceutical sector, with some biotechnology company stock prices fluctuating after the announcement.
