Pricing
What a call costs, read live from configuration and the model catalog.
Formula
Hourly settlement runsA markup burn row appears
Example dataAll integers in micro-USD, rounded up.
Providercost+Platform feen/a+Endpoint markupn/a=Paid callprice
Markupsplit→Launcher royaltyn/aBuy and burnn/a
- cost = list price of the served model × exact tokens (input, output, cache read, cache write)
- fee = cost × the platform fee
- markup = cost × endpoint markup, only for balance-paid calls to an endpoint
- royalty = the launcher share of markup, burn = the rest
- Allowance calls pay cost + fee. Raw models pay cost + fee. Tests pay cost only, from the platform budget.
Model list prices
Open the model pickerEndpoints first, then raw models
Example dataLive from GET /api/v1/models, per million tokens.
Loading Configuration
Endpoint prices
Open Explore and tap $JPXIts market and service open
Example dataAn endpoint lists priceMicroPer1M = list price × (1 + the platform fee + its markup), rounded up.