OpenAI API Pricing Calculator

This free OpenAI API pricing calculator shows what your app will cost on every current OpenAI text model. Pick a model, enter the tokens in each request and how many requests you expect, and see the cost per request, per day and per month, with prompt caching and Batch, Flex and Fast pricing included.

Under the calculator is OpenAI’s full API price list for every model in it, checked against OpenAI’s pricing page on October 18, 2026. Nothing you type leaves your browser.

How to use the OpenAI API cost calculator

1. Pick a model. The three models OpenAI lists as its latest are at the top. Models OpenAI is retiring show their shutdown date.
2. Pick a processing tier. Standard is the normal price. Batch and Flex cost half as much on most models if you can wait, and Fast costs more for quicker answers.
3. Enter tokens per request. Count everything you send (instructions, chat history, documents) as input and the reply as output. Not sure? Paste a sample and the calculator estimates tokens at about 4 characters per token.
4. Enter your traffic. Requests per day and days per month turn the cost per request into a monthly bill.
5. Add caching if part of each prompt repeats, then check the table to see the same workload on every other model.

How OpenAI API pricing works

OpenAI charges per token, priced per million tokens, with separate rates for each part of a request:

• Input: everything you send the model.
• Cached input: the start of a prompt that OpenAI processed recently, billed at a steep discount. On the latest models a cached token costs a tenth of the normal input price or less.
• Cache writes: on GPT-5.6 and later models, storing a prompt prefix in the cache costs 1.25 times the input price. Earlier models charge nothing extra for writes.
• Output: what the model writes back, including reasoning tokens, which OpenAI bills as output even though you don’t see them. On the three latest models, output costs five times the input price.
• Long prompts: prompts over 272K input tokens cost more on the models that list long context prices.

Put together: (input tokens x input price + cached tokens x cached price + output tokens x output price) / 1,000,000 = cost per request. Multiply by requests per day and days per month for the monthly bill.

OpenAI’s latest models at a glance

Model	Input	Cached input	Output	Batch input / output
gpt-6-astra	$10.00	$1.00	$50.00	$5.00 / $25.00
gpt-6.1-sol	$2.00	$0.10	$10.00	$1.00 / $5.00
gpt-6-luna	$0.10	$0.01	$0.50	$0.05 / $0.25

Standard prices per million tokens for prompts up to 272K tokens. The calculator above lists every model and tier.

Standard, Batch, Flex or Fast?

Tier	Price	Best for
Standard	List price	Chat apps and anything a person is waiting on
Batch	50% off	Big jobs that can wait up to 24 hours, like tagging or summarizing a backlog
Flex	Batch prices, on supported models	Background calls that can live with slower answers and occasional unavailability
Fast	About 2x on the latest models	Apps where speed matters more than cost

Not every model offers every tier. The calculator greys out the tiers a model doesn’t have.

Ways to cut your OpenAI API bill

• Cap the output. Output tokens cost several times more than input, so a lower max output setting and instructions like “answer in 3 sentences” save real money.
• Put the repeated part first and cache it. Long instructions and reference documents at the start of every prompt can be served from cache. On GPT-5.6 and later, caching needs at least 1,024 repeated tokens.
• Send anything that can wait through Batch. Half price for overnight jobs is the easiest saving there is.
• Test a smaller model. The comparison table shows what a mini or nano model would save. Try it on your real prompts before you switch.
• Watch shutdown dates. OpenAI plans to shut down gpt-4o, gpt-4o-mini, gpt-4.1 and gpt-4.1-mini on January 14, 2027, so budget for the move.

Prices exclude tax and any negotiated discounts. OpenAI models on Microsoft Azure and Amazon Bedrock are billed by those services at their own prices.

Related on HOW TO // AI

• Claude API cost calculator: the same calculator for Anthropic’s Claude models.
• AI token calculator: paste any text to count its tokens and see what it costs.
• ChatGPT pricing: the ChatGPT plans, which OpenAI bills separately from the API.
• Budgets and alerts for usage-based AI spend: how to keep an API bill from surprising you.

Checked against OpenAI’s API pricing pages (developers.openai.com and openai.com), prompt caching, Batch, Flex processing, reasoning and deprecations pages on October 18, 2026.

Frequently asked questions

How much does the OpenAI API cost?

It depends on the model and how many tokens you use. On October 18, 2026, OpenAI's three latest models ranged from $0.10 per million input tokens and $0.50 per million output tokens (gpt-6-luna) to $10 and $50 (gpt-6-astra), with gpt-6.1-sol at $2 and $10.

How many words is 1 million tokens?

About 750,000 English words. OpenAI's rule of thumb is that 1 token is about 4 characters or 0.75 words of English text. Other languages and code usually use more tokens.

How much does the OpenAI Batch API save?

Batch costs 50% less than Standard on the models that offer it. In return, results arrive within 24 hours instead of right away.

Is the OpenAI API included in ChatGPT Plus?

No. OpenAI bills the API separately from ChatGPT Plus, Business, Enterprise and Edu subscriptions, so API usage is paid per token on its own account.

What is cached input in OpenAI API pricing?

When the start of your prompt matches one OpenAI processed recently, those tokens are billed at the lower cached-input price. On GPT-5.6 and later, the repeated part must be at least 1,024 tokens, and the first write to the cache costs 1.25 times the normal input price.