Start here / JevModel
JevModel pricing and free Jev runs
Understand the 10,000 signup tokens, Flow usage, input-token billing, packs and rollover plans. Choose a plan after measuring your own workload.
Last updated:
At a glance
- Verified accounts get 10,000 free tokens once, about 50 short requests, and visitors can run 3 demos a day without signing up. Free access is an allowance, not open model weights.
- JevModel plans, direct model-provider rates and the cost of operating open alternatives are different budgets. Measure successful decisions and review effort before choosing.
1. Start with the signup tokens
Sign in and verify your email to receive 10,000 free tokens once. Playground, each Batch row, Decision Flows and the ordinary API all draw from this one balance; opening another tab or creating another API key does not create a second allowance. Free accounts can save one Decision Flow and upload 50-row batches. An illustrative homepage answer is not a live run.
Your first Jev decisionJevModel API: request and response2. Count a Flow differently from a normal run
Decision Flows run on your token balance, including signup tokens; free accounts can save one flow and any purchase unlocks unlimited flows. Each question actually visited on the chosen path consumes the model-reported input tokens, irrespective of the flow’s total length. A later node failure does not undo paid questions that already succeeded. Review the returned steps and balance before retrying; a flow outcome does not itself send a notification or issue a refund.
AI Decision Flows with Jev and human review3. Separate service prices from model-provider prices
JevModel service subscriptions and token packs are listed on Pricing. For example, Starter is $9.90 per month for 30M input tokens; the smallest one-time pack is $4.90 for 8M tokens. These values come from this site's checkout catalog. They are not TypeSafe or OpenRouter direct pricing. JevModel is independently operated and is not affiliated with TypeSafe AI. Confirm the current product and checkout before purchasing.
4. Estimate from measured inputs
Successful ordinary requests settle against model-reported input tokens; output is free. Token packs never expire, unused subscription tokens roll over, and yearly plans deliver all 12 months of tokens up front at 50% off. Request estimates on the pricing page assume about 200 input tokens per short question. For a workload estimate, record actual input-token usage across representative short and long states, multiply by successful request volume, and include retries and every visited Flow question. A per-row character count is not a reliable token count. Check failures and uncertain results as well as the cheapest estimate.
Evaluate a Jev workflowBatch text classification with the JevModel APIIs Jev open source?JevModel is independent and not affiliated with TypeSafe AI.