Jev AI pricing: what does it cost?

The published TypeSafe price

At this review, TypeSafe lists Jev 1.13 at $0.042 per million input tokens, equivalent to $42 per billion. Output tokens are free. This is the rate on the official model page, not a quote for every reseller or custom plan.

USD. Arithmetic at the published rate; excludes any other provider fees, taxes, or surrounding infrastructure.

Estimate a workload

Start with measured input usage per call. For a hypothetical workflow that averages 2,000 input tokens and makes 10,000 calls:

2,000 tokens × 10,000 calls = 20,000,000 input tokens
20,000,000 / 1,000,000 × $0.042 = $0.84

The 2,000-token input is an assumption, not a benchmark. Real usage changes with the context and questions. Retries and repeated calls add work, so measure actual usage during a small pilot instead of budgeting only from a sample message’s length.

A model’s invoice is also not your whole operating cost. Include the engineering time to evaluate wrong routes, the human queue, and any separate text-generation steps.

Is Jev AI free?

There is a difference between free output tokens and a free service. Vercel announced a Jev promotion on AI Gateway through September 25, 2026. That dated offer does not establish permanent free access on TypeSafe or other providers. Check Vercel’s announcement and your chosen provider’s current billing screen.

We do not promise credits, an unrestricted trial, or a particular account entitlement. Our access guide covers the official entry points, and the Gateway walkthrough explains the separate integration.

Compare costs on the same task

Our recommendation is to compare cost per acceptable outcome. Use the same messages and acceptance rules for each candidate. Include retries and review time; a cheaper wrong decision can create more work downstream.

For example, measure how many messages are correctly routed without intervention. Keep the uncertain cases visible in the comparison. The Jev versus LLM guide gives a practical evaluation checklist.