LLM API Cost Calculator

How the cost is calculated

Monthly cost = (input tokens × input price) + (output tokens × output price), with prices quoted per 1 million tokens. If you set a cached share, that portion of your input is billed at the model's cache-read price instead of the full input price. Batch mode uses each provider's asynchronous batch prices, which are typically about half the standard rate, and ignores the cache setting.

Estimating your token usage

In English, one token is roughly four characters, or about three quarters of a word. A 1,000-word prompt is therefore around 1,300 tokens. Check the usage fields in your provider's API responses for real numbers, since system prompts, chat history, and tool definitions all count as input. Many applications send far more input than they receive as output, so the input price often matters as much as the output price.

Frequently asked questions

• Why is output more expensive than input? Most providers charge several times more for generated tokens than for the tokens you send, so long answers drive cost faster than long prompts.
• What does the cached share do? Repeated prompt prefixes (such as a fixed system prompt) can be cached and re-read at a steep discount. Enter the percentage of your input you expect to hit the cache.
• Are long-context requests included? No. Some providers charge a higher rate above a context threshold. This calculator uses the standard rate only.
• How often are prices updated? The price data is refreshed weekly. The date of the last fetch is shown above. Always confirm on the provider's pricing page before budgeting.