TypeSafe AI opened Jev to all users on September 20, 2026, five days after launching the model with a waitlist. New users receive $5 in credits, equivalent to about 120 million tokens at $0.042 per million input tokens, with no charge for output tokens. Jev is designed for fast, structured decision-making rather than essay or code generation: it accepts program state and typed queries, then returns choices, rubric-based scores or yes/no probabilities with calibrated confidence. TypeSafe AI says the model responds in 70 to 500 milliseconds and has a context window of roughly 32,000 tokens. Integrations have gone live with LiteLLM, Netlify AI Gateway, Cloudflare, OpenRouter and Vercel tools. Vercel previously said about 13% of its paid customers used Jev within one day of launch, while TypeSafe AI reported a 27-task comparison showing Jev was faster and cheaper than an OpenAI model; those figures remain company-reported. Independent testing has found strong performance on narrow classification tasks but also exposed vulnerability to adversarial text inputs that can produce unintended decisions, including false negatives. No extensive production-level benchmarks or regulatory examinations have been documented since open access began.