OpenAI deploys almost two thousand agents to analyze mouse brain data

OpenAI, the drug discovery company PerturbAI, and Scripps Research announced on October 21 that they used 1,946 artificial intelligence agents to analyze biological data from millions of mouse brain cells. R&D World reported that the network of digital agents evaluated a whole-brain clustered regularly interspaced short palindromic repeats atlas, also known as CRISPR atlas, and generated research leads in 32 minutes. [1]

The research team deployed parallel instances of GPT-Rosalind, a specialized life sciences artificial intelligence model that OpenAI introduced earlier this year. According to R&D World, the team estimated that human scientists conducting a comparable analysis and literature review would require significantly more time to process the massive biological dataset. [1]

OpenAI gained prominence through its generative language models, which include GPT-3, GPT-4, GPT-5, and recent releases such as GPT-6 Astra and GPT-6.1 Sol. The private company reached a valuation of approximately 850 billion dollars in early 2026. [4]

The study demonstrates how large ensembles of specialized software agents can examine existing cellular datasets to formulate hypotheses for human scientists to test in future laboratory experiments. [1]