AI API cost calculator (OpenAI, Claude, Gemini)
Estimate what your tokens cost per request, per day and per month, and compare models.
Estimates assume about 4 characters per token for English; other languages and code use more.
Per request
$0.04
Per day
$3.50
Per month (30 days)
$105.00
OpenAI gpt-6-astra: 10 in / 50 out per 1M tokens. Month = 30 days.
All models, cheapest per month first
Prices as of 2026-11-05, check the provider's page before you budget: OpenAI, Anthropic, Google. USD per 1M tokens. Edit any price to try your own; edits stay in this tab only.
| Model | Input | Output | Cached | Per month |
|---|---|---|---|---|
| OpenAI | $1.05 | |||
| Prompts over 100K tokens: $0.50 in / $2.50 outAnthropic | $1.05 | |||
| $3.00 | ||||
| $4.65 | ||||
| Rises to $1.50 in / $7.50 out on 2027-01-01Google | $7.88 | |||
| OpenAI | $10.50 | |||
| OpenAI | $21.00 | |||
| Anthropic | $21.00 | |||
| Rises to $4.00 in / $20.00 out on 2026-12-01Google | $21.00 | |||
| Prompts over 200K tokens: $4.00 in / $18.00 outGoogle | $24.00 | |||
| Anthropic | $42.00 | |||
| OpenAI | $105.00 | |||
| Anthropic | $105.00 |
How it works
Estimate AI API cost per request, day and month from tokens and requests, and compare OpenAI, Claude and Gemini models. Edit any price. Free, nothing uploaded.
- Pick a model, then enter tokens per request and requests per day, or paste text to estimate.
- Read the cost per request, per day and per month, and scan the comparison table.
- Edit any price to match the provider's page, then copy the summary or download the CSV.
Questions
Are these prices current?
They are shown with an "as of" date and a link to each provider's pricing page. Providers change prices, so check the page before you budget, and edit any price in the table to use the current one. A small note under a model flags a coming price change or a higher rate for long prompts; the table uses the standard rate.
How is the token estimate made?
When you paste text, the tool assumes about 4 characters per token, which is a rough rule for English. Other languages and code usually use more tokens per character, so treat it as an estimate, not a bill.
What does the cached share do?
It sets what percent of your input tokens are billed at the cached-input price. Models with no listed cached price bill those tokens at the normal input price.
How is the month calculated?
The monthly cost is the daily cost times 30 days. Batch discounts, long-context surcharges, tools and image or audio tokens are not included.
Are my edited prices or pasted text saved?
No. Everything stays in this tab's memory and is gone when you reload. Nothing is sent anywhere.