What is Jev?
Jev is TypeSafe AI’s decision model: you send it text or JSON plus questions with answers you define, and it returns typed answers with probabilities instead of generating text. It opened in early access on September 15, 2026, and the current version on TypeSafe’s models page is jev-1.13.0, with 64k tokens of context per request and text as the only input.
TypeSafe was founded in 2024 by Diogo Almeida, who worked on RLHF, InstructGPT, and ChatGPT at OpenAI, with co-founders Erik Gafni and Sasha Sheng, and SiliconANGLE reports a $40 million seed round led by DCVC. The launch post says the name comes from William Stanley Jevons: more efficient steam engines raised the demand for coal, and TypeSafe expects cheaper intelligence to do the same.
Two things shape whether it fits your stack. The same weights serve every customer, so unlike a classifier you train once on your own labels, everything you can tune lives in the request. And the models page says English is where accuracy is best, so a non-English workload needs its own test before anything depends on it.