Patronus AI raises $50 million in a Series B round to build simulated environments that test how reliably AI agents complete complex tasks. Its digital world models replicate websites and internal systems, where agents are evaluated on tasks in software engineering and finance. The round is led by Greenfield Partners and brings the startup’s total funding to $70 million.
Patronus AI partners with CARIAD, Volkswagen Group's software company, to continuously enhance its in-vehicle AI assistants. The collaboration uses efficient quality checks and reliable reporting to support that work.
Patronus AI releases GLIDER, a 3.8-billion-parameter open-source model built to evaluate language models quickly and explainably. The company says GLIDER is the smallest model to outperform GPT-4o-mini as an evaluator, and describes it as supporting real-time evaluations with transparent reasoning and actionable insights.
