OpenAI API Pricing Explained: How Much Does GPT Cost?
A complete guide to OpenAI API pricing, breaking down the true costs of GPT-5.5, GPT-5.5 Mini, and prompt caching.
Understanding OpenAI's API pricing can be complicated, especially with the introduction of new tiers, cached tokens, and lightweight models.
If you are building an AI application, calculating your expected OpenAI bill is critical to ensuring your software remains profitable. Here is the complete breakdown of OpenAI's current pricing model.
1. The Flagship: GPT-5.5 Pricing
GPT-5.5 is OpenAI's most powerful reasoning model, designed for complex coding, heavy logic, and deep analysis. It is also the most expensive.
- Input Cost: $5 per 1 Million tokens.
- Output Cost: $30 per 1 Million tokens.
- Cached Input Cost: $0.5 per 1 Million tokens.
When to use it: Use GPT-5.5 only when your task requires high-level reasoning. Because output tokens cost significantly more than input tokens, avoid using this model for writing massive essays or generating long, simple text blocks.
2. The Workhorse: GPT-5.5 Mini Pricing
For 90% of standard AI tasks (summarization, categorization, basic chatting, and extraction), GPT-5.5 Mini is the undisputed champion of cost-efficiency.
- Input Cost: $0.15 per 1 Million tokens.
- Output Cost: $0.6 per 1 Million tokens.
- Cached Input Cost: $0.075 per 1 Million tokens.
When to use it: Always default to GPT-5.5 Mini unless you absolutely need the reasoning power of the flagship model. It is nearly 30x cheaper to run than its bigger sibling.
3. How Prompt Caching Saves You Money
If you send the same massive system prompt (like a set of guidelines or a large document) repeatedly within a short timeframe, OpenAI will cache those tokens on their servers.
When a prompt is cached, you receive a massive discount. For example, GPT-5.5's input cost drops from $5 to $0.5 per million tokens. Ensure your static system prompts are placed at the very beginning of your API payload to maximize this discount.
Calculate Your OpenAI Bill
Want to see exactly what your monthly bill will look like based on your expected user traffic? Use our free calculator to model your costs dynamically.