TypeSafe Launches Jev, an AI Model That Returns Probabilities Instead of Text
TypeSafe released Jev on September 15, 2026, describing it as a 'System One' AI model designed for fast, deterministic classifications rather than conversational responses. Instead of generating text, Jev accepts a state object and a list of structured questions, returning a probability score between 0 and 1 for each query in a single parallel pass. The model supports three answer primitives — noul (yes/no probability), choice (categorical selection), and score (ordered rubric) — and delivers responses in 70 to 500 milliseconds. Pricing is set at $42 per billion input tokens, with output tokens provided free of charge. One developer used Jev to automate pull request reviews, sending six simultaneous questions per request and using the returned probabilities to programmatically assign pass, review, or block verdicts.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Discussion (0)
Log in to join the discussion and vote.
Log in