Short answer
OpenAI lists GPT-5 API rates of $1.25 per 1 million input tokens, $0.125 per 1 million cached input tokens, and $10 per 1 million output tokens. These are API usage rates, not a monthly ChatGPT subscription price. OpenAI’s API documentation identifies the model as gpt-5, gives gpt-5-2025-08-07 as its default snapshot, and labels GPT-5 deprecated. It recommends GPT-6 Astra as the newer model. Check your API account’s model availability before planning a new integration around GPT-5.
OpenAI also says it retired GPT-5 Instant and Thinking from ChatGPT on February 13, 2026. A ChatGPT subscription is therefore not a way to select those GPT-5 models in the app today.
GPT-5 API prices
| API token type | Listed price (USD per 1 million tokens) |
|---|---|
| Input | $1.25 |
| Cached input | $0.125 |
| Output | $10.00 |
The rates are for the specific API model gpt-5; do not treat them as prices for GPT-5.5, GPT-5.6, or another model. OpenAI’s documentation marks GPT-5 deprecated, so these listed rates should not be taken as a guarantee that the model is available for a new production integration.
Estimate an API bill
For a billing period, use the token counts reported for your requests. If you have total input tokens and the cached-input count, calculate:
Estimated cost = ((total input tokens − cached input tokens) ÷ 1,000,000 × $1.25) + (cached input tokens ÷ 1,000,000 × $0.125) + (output tokens ÷ 1,000,000 × $10.00)
Count cached input separately from uncached input. The cached count is part of total input, not extra input; use the API’s reported cached-token usage rather than assuming repeated text qualifies for cached pricing.
For example, 2 million input tokens, none cached, plus 200,000 output tokens would cost:
- Input: 2 × $1.25 = $2.50
- Output: 0.2 × $10.00 = $2.00
- Estimated total: $4.50
This is a hypothetical calculation using the listed token rates. It excludes any separate charges for tools or other billable features. For a production estimate, add up expected requests and average input, cached-input, and output token counts over the same billing period.
Does a ChatGPT subscription include GPT-5 API usage?
No, not as standard API access. OpenAI says ChatGPT and the API platform have separate billing systems; API usage is billed separately from a ChatGPT subscription. Eligible paid Pro subscribers and ChatGPT Business workspaces may be able to claim promotional API credits, subject to eligibility and activation requirements. Those credits do not make API usage part of the ChatGPT subscription or change GPT-5’s token rates.
What to check before budgeting
- Confirm model access. GPT-5 is deprecated. Check that the exact model is available in your API account before building a new integration around it.
- Use the right price list. GPT-5’s rates are not a substitute for a successor model’s prices, or vice versa.
- Use actual token categories. Keep uncached input, cached input, and output distinct; don’t count cached tokens twice.
- Estimate your own traffic. A useful budget needs the billing period, expected request volume, and average token counts. The $4.50 example above is arithmetic, not a measured production workload.
Bottom line: OpenAI lists GPT-5 API pricing at $1.25 per million input tokens, $0.125 per million cached input tokens, and $10 per million output tokens. GPT-5 is deprecated and no longer selectable in ChatGPT, so confirm API availability before using those rates to budget a new production workload.
Sources
Prompt Caching in the API | OpenAI
Retiring GPT-4o and other ChatGPT models | OpenAI Help Center
Managing billing for ChatGPT and the API platform | OpenAI Help Center
API credits benefit for eligible ChatGPT subscriptions | OpenAI Help Center