TypeSafe opens Jev AI model to all users

TypeSafe AI opened access to Jev, its decision-focused AI model, to all users on September 20, 2026, five days after launching the model with a waitlist on September 15. New users receive $5 in credits, equivalent to roughly 120 million tokens, according to the company. Jev does not generate text. Instead, it accepts a program state alongside typed questions and returns structured outputs such as user-defined choices, rubric-based scores, or yes-no probabilities, each paired with a calibrated confidence score. TypeSafe describes it as a System One model, borrowing the term from Daniel Kahneman's framework for fast, intuitive thinking. Because every question in a request is evaluated in parallel, adding more questions barely changes response time. TypeSafe prices Jev at $0.042 per million input tokens, with output tokens free since nothing is generated to bill for. The company says results return in under half a second, with latencies between 70 and 500 milliseconds. Sources differ on the context window, reporting either roughly 32,000 tokens or a cap of 64,000 tokens.