A Thousand OpenAI Agents Sort Through a Mouse Brain Atlas in 32 Minutes
OpenAI, PerturbAI, and Scripps Research ran thousands of agents on a model called GPT-Rosalind over a mouse brain gene-activity dataset, turning up new leads for studying Alzheimer's and Parkinson's.
OpenAI, PerturbAI, and Scripps Research ran thousands of AI agents on a model called GPT-Rosalind. OpenAI is an AI company. PerturbAI is a biotech company. Scripps Research is a research institute. The agents worked through a massive dataset on gene activity in mouse brain cells. Each agent studied one knocked-out gene. It checked how the cells reacted to that, then looked for an explanation in scientific papers. The whole analysis took 32 minutes. Human researchers would have needed months, maybe a year, for the same work. In a blind test, experts picked the report built from many separate agents over the report from a single agent that had seen the entire dataset. They chose the distributed report 75% of the time.
The single agent that saw everything cited only 11 scientific papers. The thousand agents, each working on one gene, cited more than 700 papers combined. OpenAI's own report calls these results leads for further study in labs. It is clear about one thing: these are not finished answers about Alzheimer's or Parkinson's disease.
Source: openai.com
Free course
Stop reading about AI — start building with it
The free Claude Code course: your first site, tool or game — no coding. No upsells, no cross-sells — nothing to buy here.
Start free →
Author
Evgenii Arsentev
PhD · Chief Executive Officer, digital health
Articles · Latest articles