LLM API Cost Calculator

Forecast monthly AI API cost by model with editable prices, caching and batch discounts.

Cached input is usually much cheaper
Enter 1 for yes, 0 for no

Prices (USD per 1M tokens, editable)

Prices are indicative list prices in USD per million tokens as of November 2026, taken from public price comparisons. Providers change prices and have tiers (batch, cached input, long context), so edit any row to match your own pricing page.

LLM API cost calculator

Estimate what an AI feature will cost to run. Enter the average input and output tokens per request and how many requests you expect each day. The calculator ranks popular models by monthly cost and lets you edit every price, apply a prompt-caching share and switch on batch pricing.

How API pricing works

Most providers charge per million tokens, with a lower price for input and a higher price for output. Reasoning tokens usually bill as output. Cached input and batch processing can cut costs substantially, which is why those two fields are included.

Why prices are editable

Model lineups and prices change often, and some models charge more above a long-context threshold. The default table holds indicative list prices from public comparisons as of November 2026. Overwrite any figure with the number on your provider's pricing page for an accurate forecast.

Cutting your AI bill

Shorten system prompts, cache repeated context, cap output length, route simple requests to cheaper models, and use batch endpoints for work that can wait. Use the LLM Token Counter to measure real prompt size first.

What it does not include

Image, audio and tool-use charges, fine-tuning, hosting, embeddings and rate-limit tiers are not modelled. Treat the result as a planning estimate.

Preguntas frecuentes

How do I estimate cost per request?

Input tokens × input price plus output tokens × output price, each divided by one million.

Are the prices guaranteed?

No. They are indicative and editable. Check the provider's page.

How does caching change the cost?

Cached input tokens are billed at a discount. Enter your share and discount to see the effect.

What is the batch discount?

Some providers charge about half price for asynchronous batch jobs.

Why is output more expensive?

Providers price generated tokens higher than input tokens.

Más en AI Tools