OpenAI and longevity start-up Retro Biosciences said an experimental large language model dubbed “GPT-4b micro” has designed new versions of the Nobel-winning Yamanaka factors that reprogram adult cells into induced pluripotent stem cells. Lab tests show the AI-derived proteins increased reprogramming efficiency more than 50-fold compared with the standard OSKM cocktail, the companies told reporters. The collaboration began about two and a half years ago and has been running GPT-4b micro in Retro’s wet-lab pipelines since January. While the model itself will not be released publicly, the groups said the results illustrate how generative AI could speed discovery in protein engineering and regenerative medicine. Retro Biosciences, backed with a $180 million investment from OpenAI chief executive officer Sam Altman in 2022, aims to add roughly a decade to healthy human lifespan. Executives said the new data support moving the redesigned Yamanaka factors toward further pre-clinical studies, though no timelines for clinical testing were disclosed.