32 Minutes: Thousands Of AI Agents Read Millions Of Brain Cells
OpenAI published this on October 21, 2026, with PerturbAI and Scripps Research. Everything below comes from that first-party post.
Transcript
Thousands of AI agents analyzed brain data today. OpenAI says it took thirty two minutes.
OpenAI did this with PerturbAI and Scripps Research, on existing data from millions of mouse brain cells.
OpenAI calls the approach distributed biological reasoning, where each agent investigated one gene disruption and proposed an explanation from published research.
In blinded comparisons, human experts preferred the distributed agent reports seventy five percent of the time over a single agent.
These are leads for follow up experiments, not confirmed findings. The atlas and the preference judgments are OpenAI's own, and no treatment or clinical result is claimed.
Inboxsmith builds AI receptionists for small business. Please like and subscribe for more news.
Sources
Every claim in this video comes from the top ranking coverage of this topic. The claims and where each one came from:
- Working together, OpenAI, PerturbAI, and Scripps Research used thousands of AI agents powered by GPT-Rosalind to analyze and interpret existing experimental data from millions of mouse brain cells in just 32 minutes.(OpenAI (openai.com), first party publication titled "Using thousands of AI agents to uncover new leads in biological data", October 21, 2026)
- The team divided the analysis among thousands of AI agents powered by GPT-Rosalind, an approach we call distributed biological reasoning. The agents examined a whole-brain CRISPR atlas, a dataset showing how gene disruptions affect different types of cells across the mouse brain.(OpenAI (openai.com), first party publication titled "Using thousands of AI agents to uncover new leads in biological data", October 21, 2026)
- Each agent investigated a gene disruption, analyzed cellular responses, searched scientific literature, and proposed explanations for the results. Their reports were combined into a reasoning atlas linking genes, cellular effects, relevant publications, and possible biological mechanisms.(OpenAI (openai.com), first party publication titled "Using thousands of AI agents to uncover new leads in biological data", October 21, 2026)
- In blinded comparisons, human experts preferred distributed-agent reports over single-agent reports 75% of the time.(OpenAI (openai.com), first party publication titled "Using thousands of AI agents to uncover new leads in biological data", October 21, 2026)
- As the workload grew, the single agent searched less of the literature, citing only 11 references at the largest workload, while the distributed agents cited more than 700 references.(OpenAI (openai.com), first party publication titled "Using thousands of AI agents to uncover new leads in biological data", October 21, 2026)
- The article describes research leads and hypotheses for follow-up experiments, not confirmed findings: "These findings help prioritize further experiments needed to test the proposed biological mechanisms" and "This raised a question for follow-up experiments". The paper, the reasoning atlas and the blinded expert preference judgments are the work of OpenAI and its partners PerturbAI and Scripps Research. No treatment, no clinical result and no timeline is claimed.(OpenAI (openai.com), first party publication titled "Using thousands of AI agents to uncover new leads in biological data", October 21, 2026)