Usage & billing
Practical guidance for using NeonAPI.
Read the published price
The model page is the starting point for current rates. Pay attention to currency, token unit, group eligibility, context tiers, cache categories and per-request or per-task units. Featured comparisons are reference tariffs, not a promise that every request receives the same discount.
Understand usage
Input and output tokens may have different rates. Some models distinguish cache reads and cache creation. Image, audio or asynchronous video tasks can use other billing dimensions. Use the specific model page rather than a universal token formula.
Review a charge
- Find the request in usage logs by time and model.
- Check the reported usage, group and billing conditions.
- Compare the applicable model tariff, including any tier or task dimensions.
- If a difference remains, provide a redacted request identifier to support.
Balance, plans and pending tasks
Use only the payment and plan methods actually shown by the console. Do not assume a particular payment currency, refund term or expiry. Asynchronous tasks may reserve quota before completion; inspect the final task status and usage record before retrying.
A failed client connection does not prove the upstream request was never processed. Repeating a generation request may create a second billable request.