source: techcrunch ai: prentis, new ai lab co-founded by reid hoffman, mark pincus in talks to raise $100m
level: business
prentis, a new ai research lab, is training models to learn how office workers handle everyday tasks across documents and systems. the goal is to create ai agents that can control computers to automate these workflows. the startup has already signed contracts worth up to $50 million with customers in healthcare, manufacturing, and goods production. investor materials predict an annualized run rate of $75 million by the third quarter of this year, though those figures depend on performance and realized savings.
the company claims its hive-32b model outperforms larger rivals like openai's gpt-5.4 and anthropic's claude opus 4.6 on two computer-use benchmarks: windowsagentarena and screenspot-v2. prentis says its smaller model is about ten times cheaper per task than frontier apis, making it more economical for widespread deployment. however, these benchmark results have not been independently verified.
prentis enters a crowded field where anthropic, openai, and thinking machines lab are also developing computer-use agents. anthropic recently acquired vercept, a seattle startup in the same space. prentis has hired over 25 employees, including researchers from openai, google deepmind, meta, tencent, and alibaba. the lab is part of titan, a holding company founded by ceo ritankar das that builds and operates ai businesses.
why it matters: automating routine office tasks with smaller, cheaper models could make ai agents practical for everyday business use, potentially reducing costs and freeing workers for higher-value work.
source: techcrunch ai: prentis, new ai lab co-founded by reid hoffman, mark pincus in talks to raise $100m