AI Inference Billing

Billing Logic	Minimum billable usage	Time to charge
Based on usage volume (number of tokens/characters)	1 token	Every 5 minutes

Billing Equation:

• Usage tokens: Input tokens + Output tokens

• Cost: Input tokens × input price / 1M + Output tokens × output price / 1M

Example:

You use model Qwen3-32B priced at $0.165/1M input tokens and $0.187/1M output tokens, using 13,394 input tokens and 127 output tokens

• Usage token: 13,394 input tokens and 127 output tokens

• Cost = (13,394 × 0.165 + 127 × 0.187) / 1,000,000 = $0.0022