OpenAI API pricing: what every GPT model costs per 1M tokens
Input, cached input, cache-write and output prices for OpenAI’s current models, the long-context surcharge, Batch, Flex and Fast rates, and a worked monthly cost example.
Last updated October 22, 2026
How much does the OpenAI API cost?
On October 22, 2026, OpenAI’s flagship gpt-6-astra cost $10 per 1M input tokens and $50 per 1M output tokens; gpt-6.1-sol cost $2 and $10; gpt-6-luna cost $0.10 and $0.50. Prompts over 272K tokens cost more. Batch and Flex halve these prices, and cached input costs a tenth or less.
What does each OpenAI model cost per 1M tokens?
These are the standard prices for the models OpenAI lists under “Flagship models” on its API pricing page. Short context means up to 272K input tokens; long context means more than 272K. OpenAI’s models page recommends gpt-6-astra for complex reasoning and coding, gpt-6.1-sol to balance intelligence and cost, and gpt-6-luna for cost-sensitive, high-volume work.
| Model | Input | Cached input | 30-min cache write | Output | Long context (>272K) input / output |
|---|---|---|---|---|---|
| gpt-6-astra | $10.00 | $1.00 | $12.50 | $50.00 | $20.00 / $75.00 |
| gpt-6.1-sol | $2.00 | $0.10 | $2.50 | $10.00 | $4.00 / $15.00 |
| gpt-6-sol | $2.00 | $0.20 | $2.50 | $10.00 | $4.00 / $15.00 |
| gpt-6-luna | $0.10 | $0.01 | $0.125 | $0.50 | $0.20 / $0.75 |
| gpt-5.6-sol | $4.00 | $0.40 | $5.00 | $20.00 | $8.00 / $30.00 |
| gpt-5.6-terra | $2.00 | $0.20 | $2.50 | $12.00 | $4.00 / $18.00 |
| gpt-5.6-luna | $0.20 | $0.02 | $0.25 | $1.20 | $0.40 / $1.80 |
| gpt-5.5 | $5.00 | $0.50 | — | $30.00 | $10.00 / $45.00 |
| gpt-5.5-pro | $30.00 | — | — | $180.00 | $60.00 / $270.00 |
| gpt-5.4 | $2.50 | $0.25 | — | $15.00 | $5.00 / $22.50 |
| gpt-5.4-pro | $30.00 | — | — | $180.00 | $60.00 / $270.00 |
Prices checked October 22, 2026. Providers change prices often; confirm on OpenAI’s API pricing page.
The page also lists a 12-hour cache write for the GPT-6 and GPT-5.6 models at twice the 30-minute write price (for example $25.00 on gpt-6-astra). It says GPT-5.6 Sol’s promotional pricing runs at least through November 21, 2026, and that regional processing (data residency) endpoints cost 10% more for models released on or after March 5, 2026. Older models such as gpt-5.4-mini, gpt-4.1 and o3 are still priced further down the same page.
How does OpenAI API billing work?
- You pay per token. Input tokens (your instructions, conversation history, tool definitions and any retrieved content) and output tokens (the reply) are counted separately and billed at the model’s rates.
- Reasoning tokens are output. OpenAI’s reasoning guide says reasoning tokens are billed as output tokens, even though they are not shown in the reply.
- Prompt length can change the rate. On the models in the table, a request with more than 272K input tokens is billed at the long-context prices.
- Caching changes the input rate. According to OpenAI’s prompt caching guide, each input token is billed at the uncached, cached-input or cache-write rate; for GPT-5.6 and later, writes cost 1.25× input for the default 30-minute lifetime or 2.5× for 12 hours, and reads cost 0.1× (0.05× on gpt-6.1-sol).
- Tools cost extra. Web search is $10.00 per 1,000 calls plus search content tokens at model rates; file search tool calls are $2.50 per 1,000; hosted containers are priced per 20-minute session.
What would 1M requests a month cost on OpenAI?
Take 1,000,000 requests a month, each with 1,000 input tokens and 300 output tokens: 1,000M input tokens and 300M output tokens, all under the 272K long-context threshold and with no caching.
Monthly cost = 1,000 × input price + 300 × output price (prices per 1M tokens)
| Model and tier | Calculation | Monthly cost |
|---|---|---|
| gpt-6-astra, standard | 1,000 × $10.00 + 300 × $50.00 | $25,000 |
| gpt-6-astra, Batch or Flex | 1,000 × $5.00 + 300 × $25.00 | $12,500 |
| gpt-6.1-sol, standard | 1,000 × $2.00 + 300 × $10.00 | $5,000 |
| gpt-6-luna, standard | 1,000 × $0.10 + 300 × $0.50 | $250 |
Output is 300 of every 1,300 tokens here but more than half the bill on every row, because output costs five times input on these models. Count your real prompt sizes with the token calculator before you rely on an estimate.
How can you cut your OpenAI API bill?
OpenAI documents several ways to pay less. Here is how they change the price of the three featured models:
| Model | Standard | Batch | Flex | Fast |
|---|---|---|---|---|
| gpt-6-astra | $10.00 / $50.00 | $5.00 / $25.00 | $5.00 / $25.00 | $20.00 / $100.00 |
| gpt-6.1-sol | $2.00 / $10.00 | $1.00 / $5.00 | $1.00 / $5.00 | $4.00 / $20.00 |
| gpt-6-luna | $0.10 / $0.50 | $0.05 / $0.25 | $0.05 / $0.25 | $0.20 / $1.00 |
Prices checked October 22, 2026. Providers change prices often; confirm on OpenAI’s API pricing page.
- Batch API. OpenAI’s Batch guide describes a 50% discount for jobs that can wait, with each batch completing within 24 hours.
- Flex processing. The Flex guide prices requests at Batch rates in exchange for slower responses and occasional unavailability.
- Prompt caching. Put stable content (instructions, tool definitions, reference text) at the start of the prompt so it can be read from cache at a tenth of the input price or less. OpenAI’s caching guide gives the minimum cacheable length as 1,024 tokens for GPT-5.6 and later.
- Smaller models. gpt-6-luna costs 1% of gpt-6-astra per token. Send simple classification, extraction and routing requests to the smaller model and keep the flagship for hard tasks.
To estimate what caching and routing could save on your own traffic, try the LLM gateway savings calculator.
How does Perimattic LLM Gateway help with OpenAI costs?
Perimattic LLM Gateway sits between your application and model providers. Its features:
- One OpenAI-compatible API in front of OpenAI, Anthropic, Google, Mistral, DeepSeek and open models.
- Routing rules that send each request to a model by cost, latency or task, with automatic fallback when a provider errors or rate-limits.
- Exact-match and semantic caching of responses, on top of OpenAI’s own prompt caching.
- Budgets and rate limits per team, API key or customer, with cost tracking and alerts.
Get started: Sign up and our team will walk you through Perimattic LLM Gateway for your workloads.
Where do these OpenAI prices come from?
All prices on this page come from OpenAI’s API pricing page, with billing rules from OpenAI’s prompt caching, Batch and Flex guides. They are in US dollars, exclude tax, and are OpenAI’s direct API prices; OpenAI models on Amazon Bedrock and Microsoft Azure are billed by those services at their own prices. Prices checked October 22, 2026; providers change prices often — confirm on the provider’s page before you rely on them. Perimattic is not affiliated with OpenAI.
Compare OpenAI with other providers
OpenAI API pricing, answered
Is there a monthly fee for the OpenAI API?
OpenAI’s API pricing page lists usage prices per 1M tokens, plus per-call or per-session prices for tools such as web search and containers; it does not list a monthly subscription for the API. The Responses, Chat Completions, Realtime, Batch and Assistants APIs are not priced separately: tokens are billed at the chosen model’s rates.
What is the difference between Batch and Flex on OpenAI?
Both are priced at Batch API rates, half the standard price. Batch takes a file of requests and completes them within 24 hours. Flex works on normal Responses or Chat Completions calls but accepts slower responses and occasional resource unavailability. Use Batch for offline jobs and Flex for lower-priority requests that still run one at a time.
Do reasoning tokens count as output?
Yes. OpenAI’s reasoning guide says reasoning tokens are billed as output tokens, so a request that thinks for a long time can cost far more than its visible answer suggests. Check the usage object in each response and set a reasoning effort that matches the task.
Want one API for every model, with costs under control?
Sign up to tell us which models and providers you use, and our team will get in touch to set up the LLM Gateway for your stack.
Sign up