How the calculation works
Cost = (input tokens × input rate + cached tokens × cached rate + output tokens × output rate) ÷ 1,000,000.
- Separate uncached input, cached input and output token volumes.
- Multiply each token volume by its matching price per one million tokens.
- Add the three components for the cost per request.
- Multiply by expected monthly requests; inspect the component breakdown before budgeting.
Worked example
For 8,000 uncached input tokens, 2,000 cached tokens, 1,500 output tokens and 10,000 monthly requests at GPT-5.6 Luna rates, the estimate is USD 34.40 per month.
Assumptions
- Prices are in USD per one million text tokens.
- The request mix is represented by one average request.
- Tool-call, image, audio, batch, long-context and cache-write charges are excluded unless entered into the editable rates.
Limitations and responsible use
- Provider billing rules and prices can change without notice.
- Some models apply thresholds or multipliers for long prompts.
- Estimates do not include taxes, credits, enterprise discounts or infrastructure costs.
Sources and data checks
Source links were reviewed on 4 September 2026. Recheck time-sensitive rates and rules before relying on the result.