TypeSafe AI Launches Jev, a Decision Model That Costs 100x Less Than Frontier LLMs

TypeSafe AI, founded by OpenAI's Diogo Almeida, has released Jev, a "System One" model designed to make structured decisions at a fraction of the cost and latency of traditional large language models. The model, which emerged from stealth with a $40 million seed round led by DCVC, represents a fundamentally different approach to AI inference: instead of generating text token-by-token, Jev evaluates state against typed questions and returns calibrated probability scores—all in 70 to 500 milliseconds.