How Pricing Works

Qubax bills per token, not per request or per month. You load credits with crypto, and each API call deducts the precise cost down to the micro-cent. There are no subscriptions, no seat fees, and no minimum spend beyond the per-token math.

Per-token billing

Every response is priced by the number of tokens consumed — both the prompt tokens you send and the completion tokens the model generates. The unit is the micro(one-millionth of a US dollar, i.e. $0.000001). One million tokens at $3/M costs exactly 3,000,000 micros.

1 micro     = $0.000001
1 token     = price-per-million / 1,000,000  (in micros)
1,000,000 tokens @ $3/M = 3,000,000 micros = $3.00

Each completion response includes a usageobject showing prompt, completion, and total token counts, so you can track exactly what each call cost.

{
  "usage": {
    "prompt_tokens": 42,
    "completion_tokens": 128,
    "total_tokens": 170
  }
}

Credits

Your balance is a pool of prepaid credits denominated in US dollars. Every API call deducts the micro-cost from your credit balance in real time. When the balance reaches zero, requests return a 402 Payment Required(surfaced as a rate-limit error in the SDKs) until you top up.

How discounts work

Qubax negotiates wholesale rates and passes the savings on as a discount against the published retail benchmark price for each model. Discounts run up to 99% off retaildepending on the model and current wholesale capacity.

Retail benchmark (per model):  $X per million tokens
Qubax price:                  up to 99% off $X

You pay: $X * (1 - discount)
Example: $10/M retail, 90% off  ->  $1.00/M  ->  1,000,000 micros/M

The effective per-token price is always shown in the dashboard next to each model, so you know the exact rate before you send a request.

Check your balance

Query your remaining balance and usage at any time with the billing endpoint. The response returns your credit balance in micros and USD.

curl https://api.qubax.ai/v1/billing/balance \
  -H "Authorization: Bearer $QUBAX_API_KEY"

{
  "balance_micros": 15000000,
  "balance_usd": 15.0,
  "currency": "USD"
}

You can also poll this endpoint from your application to alert users when their balance runs low, or to gate expensive agent runs behind a minimum balance threshold.