Billing

Prepaid Credit Model

LLMAI operates on a pay-as-you-go basis. There are no recurring charges, no seat fees, and no usage minimums. You deposit credit into your account; that credit is drawn down as you make API calls.

The flow is straightforward:

1. Deposit credit — Add funds to your account from console.llmai.dev/billing
2. Call the API — Each request deducts the token cost (input tokens + output tokens) at the model's published rate
3. Track spend — The dashboard shows live balance, per-request cost breakdowns, and model-level usage aggregates
4. Reload when needed — Your balance carries forward indefinitely; top up whenever it runs low

How Deductions Are Calculated

Token cost is calculated separately for input and output, then summed:

request_cost = (input_tokens  / 1,000,000) × input_rate
             + (output_tokens / 1,000,000) × output_rate

Worked example — calling gpt-5.4 with a 500-token prompt that generates a 200-token reply:

cost = (500 / 1,000,000) × $0.875
     + (200 / 1,000,000) × $5.250
     = $0.0004375 + $0.0010500
     = $0.0014875

Every API response includes a usage field with the exact token counts so you can verify the deduction:

{
  "usage": {
    "prompt_tokens": 500,
    "completion_tokens": 200,
    "total_tokens": 700
  }
}

Top-Up Amounts

Need an amount outside these tiers, or a custom arrangement? Submit a request via the contact form.

Checking Your Current Balance

From the console: Billing → Balance shows your remaining credit in real time.

Via the API:

curl https://api.llmai.dev/v1/dashboard/billing/credit_grants \
  -H "Authorization: Bearer YOUR_API_KEY"

Zero-Balance Behaviour

When your balance is fully depleted, the next API request returns:

HTTP 402 Payment Required

The error body will indicate insufficient funds. Your API key stays valid — it will resume functioning immediately after you add more credit. No action beyond a top-up is required.

Refund and Adjustment Policy

Credit purchases are non-refundable. We recommend starting with a smaller deposit to validate your integration before committing to a larger amount.

If a technical fault on our end caused an incorrect deduction, use the contact form with the affected request IDs and timing. We review each case individually.