What happens for very long OpenAI prompts?

For the GPT-5.6 family, the calculator applies the provider-published long-context multiplier above 272K input tokens.

● Runs locally in your browser

AI API Cost Calculator

Q: How is monthly API cost calculated?

Per-request input and output token cost is multiplied by requests per day and 30 days.

Q: Are caching and batch discounts included?

No. The default comparison uses standard published text-token rates and excludes special discounts or tool fees.

Estimate per-request, monthly, and annual AI API spend from token volume. Compare leading LLMs with the same workload assumptions.

Input tokens per request

Output tokens per request

Requests per day

30-day requests3,000

Lowest 30-day estimate—

Standard rates: text token pricing.Excluded: caching, batch, tools and taxes.Pricing data verified —

Comparison

LLM cost by model

Change any input above and the estimates recalculate instantly in your browser.

Model	Input	Output	Per request	30 days	1 year

Informational estimate only. Provider pricing can include caching, batch processing, tool calls, regional pricing, and other rules not represented here.

Formula

How the calculator works.

Input cost = input tokens ÷ 1,000,000 × published input rate. Output cost uses the same formula with the published output rate.

Monthly and annual estimates simply scale the per-request result by request volume.

Model a real workloadUse an average request rather than a single unusually large prompt.

Separate input from outputOutput tokens are often priced materially higher than input tokens.

Check fit tooUse the Context Window Checker before assuming a very large request is valid.

FAQ

LLM API cost questions

How is monthly API cost calculated?

The per-request token estimate is multiplied by your requests per day and 30 days.

Are caching and batch discounts included?

No. V2 compares standard text-token rates so the baseline stays simple and comparable.

Why can output dominate cost?

Many providers price generated tokens higher than input tokens, so long answers can become the main cost driver.