Jev Pricing: Dated List Rate and Cost per Accepted Decision
Evaluate Jev costs from a dated rate, complete request sizes and the cost of failures. This page is a purchasing framework, not a quote for a current account.
What the launch source quoted
The September 15, 2026 TypeSafe launch post quoted $0.042 per million input tokens and free output. That is a dated vendor list rate, not a guarantee of the terms in your current checkout or another provider’s routing service. Confirm the billing source and current contract before budgeting.
Transparent arithmetic at that rate
| Hypothetical input volume | Input charge at the quoted rate |
|---|---|
| 1 million tokens | $0.042 |
| 10 million tokens | $0.42 |
| 1 billion tokens | $42 |
These calculations use the dated launch figure only. They exclude tax, minimum commitments, hosting, integration, preprocessing and human review. No production invoice was observed.
Count the entire request
A short item can still be accompanied by substantial instructions and option descriptions. Repeating those criteria across calls can dominate the token count. Record measured usage returned by the actual route and confirm how repeated questions or batches are metered. Do not price only the visible customer sentence.
The denominator that matters
Cost per accepted decision is total workload cost divided by the number of decisions meeting your acceptance criteria. A cheap wrong route can create an expensive correction. An abstention may be desirable but still requires a fallback whose cost belongs in the budget.
Compare operational options
| Option | Cost to include |
|---|---|
| Direct provider | Request usage, integration and support arrangement. |
| A routed provider | Its actual quote, availability and additional routing behavior. |
| A constrained generative model | Input and output usage, retries and any explanation step. |
| Rules or a trained classifier | Development, maintenance and monitoring alongside runtime. |
Purchase checklist
- Identify the exact model/provider route, currency and pricing date.
- Capture request-size distribution rather than one favorable example.
- Measure failure, review and fallback costs on representative traffic.
- Compare monthly expected usage with a conservative peak scenario.
- Set a budget alert and recheck the rate before expanding volume.
Download the blank decision/model evaluation worksheet. It contains fields for expected outcomes, wrong decisions, abstentions, cost and latency; no product results are prefilled.
Sources and check scope
Public-source review checked October 19, 2026. No original model benchmark or paid-account test is claimed. Product access and prices can differ by route and account.