GPT-4o API price at a glance
OpenAI’s GPT-4o model page lists $2.50 per 1 million input tokens and $10 per 1 million output tokens, in USD. API charges depend on tokens processed—not a flat price per request. OpenAI’s page also labels GPT-4o deprecated, so treat this as pricing for a workload that uses GPT-4o, and check model availability and migration requirements before choosing it for a new production system. (developers.openai.com)
| GPT-4o API usage | Price |
|---|---|
| Input tokens | $2.50 per 1M tokens |
| Output tokens | $10.00 per 1M tokens |
These are OpenAI API rates, not ChatGPT subscription prices or a third-party provider’s offer. The model page says GPT-4o accepts text and image inputs and produces text outputs; image input is therefore relevant to API usage, but the actual estimate depends on the tokens counted for your requests. (developers.openai.com)
How to estimate GPT-4o API costs
Use the following formula with token totals for the period you want to budget:
Estimated cost = (input tokens ÷ 1,000,000 × $2.50) + (output tokens ÷ 1,000,000 × $10.00)
For example, 100,000 input tokens cost $0.25, and 10,000 output tokens cost $0.10. Together, that workload would cost about $0.35 at the listed rates.
For a monthly estimate, suppose an application makes 100,000 requests per month, averaging 2,000 input tokens and 500 output tokens per request:
- Input: 100,000 × 2,000 = 200 million tokens; 200 × $2.50 = $500.
- Output: 100,000 × 500 = 50 million tokens; 50 × $10 = $500.
- Estimated total: $1,000 per month.
This is a hypothetical estimate, not a usage test or a guaranteed bill. Replace the assumptions with representative token counts from your application. Requests that include images can change input usage; don’t estimate them from text length alone.
What to check before budgeting for production
- Model status: OpenAI marks GPT-4o as deprecated on its model page. If you’re starting a new integration, verify availability and any migration guidance before building around it. (
developers.openai.com) - Actual token totals: Base the estimate on measured input and output tokens for your workload, then multiply by expected request volume. Request count alone isn’t enough to predict spend.
- Provider and model: These rates are for GPT-4o on the OpenAI API. Azure or a third-party API provider may have different pricing or terms; don’t substitute one provider’s rate for another’s.
- API versus ChatGPT: A ChatGPT subscription and API usage are separate purchasing channels. A subscription price does not tell you what a production API workload will cost.
For pricing decisions, use OpenAI’s model page for the listed GPT-4o rate and status, and check the applicable provider’s current terms before committing a production budget. (developers.openai.com)