How to use
- Enter the typical input and output tokens of one request — or open “estimate from text” and paste a real prompt to count it.
- Set how much of the input is served from the prompt cache and how many requests you expect per day.
- The table sorts every model from cheapest to most expensive and shows the cost per request, per day and per 30-day month.
- Filter by provider to compare models from one vendor. Red context numbers mean the request would not fit.
FAQ
Where do the prices come from?
From each provider’s official pricing page, standard (non-batch) tier, in USD per million tokens. The date they were checked is shown under the table. Prices change often, so confirm before committing to a budget.
What is the cached-input share?
Most providers charge much less for input tokens that hit the prompt cache — typically a long, unchanging system prompt or document. If 80% of each prompt is a reused prefix, set the slider to 80%. Cache-write surcharges are not included.
Why does the price jump for some models on long prompts?
Some models (for example Claude Haiku 5.5 above 100K input tokens and Gemini 3.1 Pro above 200K) bill the whole request at a higher rate once the prompt passes a threshold. The calculator applies that tier automatically and labels it.
Are batch or off-peak discounts included?
No. Batch APIs are usually about 50% cheaper, and DeepSeek is 50% off outside peak hours. Halve the figures for a rough estimate.