An artificial intelligence model played a role in advancing real biomedical research, according to a blog post from OpenAI.
The company says GPT-5 Pro, a version of its latest AI system, helped immunologist Derya Unutmaz make progress on a question that had stumped his lab for three years. According to OpenAI, the model offered insights into the behavior of T cells — the immune system cells that help the body fight disease.
OpenAI frames the episode as more than a lab curiosity. The company says the breakthrough could support research into cancer and autoimmune conditions, two areas where understanding how immune cells act is central to developing treatments.
The account, also surfaced through Google News, positions the work as an example of AI acting as a research collaborator rather than just a writing or coding tool. Instead of generating text, the system is described as contributing ideas that helped a domain expert interpret a long-standing scientific mystery.
It is worth noting the limits of what these sources establish. The details available come from OpenAI itself, the company that builds the model, and describe the assist in broad terms. The sources do not lay out the specific experiments, independent verification, or peer-reviewed publication behind the claim.
Still, the story matters because it is a concrete, named example of a leading AI system being credited with helping a working scientist move a stalled project forward — a sign of how these tools may increasingly shape the pace of medical discovery.