OpenAI API pricing per million tokens

How much does the OpenAI API cost per million tokens?

Each model has its own rate per million tokens for input, cached input and output. The OpenAI pricing page also splits rates by request size, into short context (272K input tokens or fewer) and long context (more than 272K input tokens).

Model	Input	Cached input	Output
gpt-6-astra	$10.00	$1.00	$50.00
gpt-6.1-sol	$2.00	$0.10	$10.00
gpt-6.1-terra	$1.00	$0.05	$5.00
gpt-6-sol	$2.00	$0.20	$10.00
gpt-6-luna	$0.10	$0.01	$0.50
gpt-5.6-sol	$4.00	$0.40	$20.00
gpt-5.6-terra	$2.00	$0.20	$12.00
gpt-5.6-luna	$0.20	$0.02	$1.20

The page says GPT-5.6 Sol’s promotional pricing is available at least through November 21, 2026. Older models, such as gpt-5.5, gpt-5.4 and gpt-4.1, are listed on the same page.

What does long context cost?

Requests over 272K input tokens cost more per token. For every model in the table, input and cached input double and output goes up by half, going by the long context rates on the same page.

Model	Input	Cached input	Output
gpt-6-astra	$20.00	$2.00	$75.00
gpt-6.1-sol	$4.00	$0.20	$15.00
gpt-6.1-terra	$2.00	$0.10	$7.50
gpt-6-sol	$4.00	$0.40	$15.00
gpt-6-luna	$0.20	$0.02	$0.75
gpt-5.6-sol	$8.00	$0.80	$30.00
gpt-5.6-terra	$4.00	$0.40	$18.00
gpt-5.6-luna	$0.40	$0.04	$1.80

How does prompt caching change the price?

Reused prompt prefixes are billed at the cached input rate, which OpenAI’s prompt caching guide says is “discounted up to 95%.” Caching is on by default for supported models.

On GPT-5.6 and later, writing a prefix to the cache has its own rate. The default 30-minute write costs 1.25x the uncached input rate, per the same guide, and a 12-hour write, set in the Responses API, costs 2.5x. Those are total write rates. OpenAI notes they aren’t surcharges on top of ordinary input. A prefix must be at least 1,024 tokens to be cached on these models.

Model	30-minute write	12-hour write
gpt-6-astra	$12.50	$25.00
gpt-6.1-sol	$2.50	$5.00
gpt-6.1-terra	$1.25	$2.50
gpt-6-sol	$2.50	$5.00
gpt-6-luna	$0.125	$0.25
gpt-5.6-sol	$5.00	$10.00
gpt-5.6-terra	$2.50	$5.00
gpt-5.6-luna	$0.25	$0.50

How much does the Batch API save?

Half. OpenAI describes the Batch API as a “50% cost discount compared to synchronous APIs.” The trade is time, since each batch completes within 24 hours. Batch jobs also use a separate pool of rate limits.

Model	Input	Cached input	Output
gpt-6-astra	$5.00	$0.50	$25.00
gpt-6.1-sol	$1.00	$0.05	$5.00
gpt-6.1-terra	$0.50	$0.025	$2.50
gpt-6-sol	$1.00	$0.10	$5.00
gpt-6-luna	$0.05	$0.005	$0.25
gpt-5.6-sol	$2.00	$0.20	$10.00
gpt-5.6-terra	$1.00	$0.10	$6.00
gpt-5.6-luna	$0.10	$0.01	$0.60

What does Flex processing cost?

The same as Batch for the models above. In exchange you get slower responses and occasional resource unavailability. OpenAI’s flex processing guide says “Tokens are priced at Batch API rates, with additional discounts from prompt caching.” The Flex table on the pricing page matches the Batch table above for these eight models.

Flex is in beta with limited model availability. When Flex lacks capacity it returns a 429 Resource Unavailable error, and OpenAI says you are not charged for it.

What does Fast mode cost?

Twice the standard rate, for every model in the table below. Fast mode used to be called Priority processing. The pricing page notes that Priority processing was renamed Fast mode on July 30, 2026.

Model	Input	Cached input	Output
gpt-6-astra	$20.00	$2.00	$100.00
gpt-6.1-sol	$4.00	$0.20	$20.00
gpt-6.1-terra	$2.00	$0.10	$10.00
gpt-6-sol	$4.00	$0.40	$20.00
gpt-6-luna	$0.20	$0.02	$1.00
gpt-5.6-sol	$8.00	$0.80	$40.00
gpt-5.6-terra	$4.00	$0.40	$24.00
gpt-5.6-luna	$0.40	$0.04	$2.40

The page also lists an Ultrafast tier for two models, and both work out to six times the standard rate. gpt-6-astra is $60.00 input, $6.00 cached input and $300.00 output per 1M tokens. gpt-6.1-sol is $12.00 input, $0.60 cached input and $60.00 output.

Are there other surcharges?

On some endpoints. Per the pricing page, regional processing (data residency) endpoints are charged a 10% uplift for models released on or after March 5, 2026. FedRAMP endpoints carry a 10% uplift as well.

How do you pay for OpenAI API usage?

Most accounts pay in advance. OpenAI’s prepaid billing article says new API accounts use prepaid billing, and purchased credits expire after 1 year. Invoices, tiers and spend limits are covered in how OpenAI and Anthropic API billing works.

To turn these rates into a monthly figure, try the AI API cost calculator. To lower the bill, read how to cut LLM API costs or how to cut OpenAI API costs on the ernd blog.