OpenAI API Cost Calculator
Estimate your monthly OpenAI API spend with this interactive calculator. Select your model, configure your usage parameters, and toggle cost-saving features like Batch API and prompt caching to see how they affect your bill. All calculations run locally in your browser.
Quick presets
At this usage level, the API costs more than a Plus subscription ($20.00/mo).
How to Estimate Your Costs
Choose Your Model
Start with the model that best matches your quality requirements. GPT-5-mini is the best balance of cost and capability for most use cases. Use GPT-5 for tasks requiring maximum quality. Use GPT-4o-mini for the absolute lowest cost on simple tasks. The model selection has the biggest impact on your overall costs.
Set Your Usage
Configure the average input tokens per request, output tokens per response, and requests per day. If you are unsure, use the quick presets (Chatbot, Content Gen, etc.) as starting points. A typical chatbot message is 200-500 input tokens with 100-300 output tokens. Content generation uses more: 1,000-3,000 input and 2,000-5,000 output.
Toggle Discounts
Enable Batch API for a 50% discount if your workload does not require real-time responses. Enable prompt caching if your requests share a common prefix (system prompt or context). Many real-world applications qualify for both discounts. The calculator shows the combined effect so you can see your optimised cost compared to baseline pricing.
API vs Subscription: Break-Even Analysis
The break-even point between the API and a ChatGPT subscription depends on your usage pattern and model choice. Here are the approximate break-even points for interactive use (where each "message" averages 500 input tokens and 300 output tokens):
| Model | Cost per "message" | Messages to hit $20 (Plus) | Messages to hit $200 (Pro) |
|---|---|---|---|
| GPT-5 | $0.0036 | ~5,500 | ~55,000 |
| GPT-5-mini | $0.0007 | ~28,500 | ~285,000 |
| GPT-4o-mini | $0.0003 | ~66,000 | ~660,000 |
Note: ChatGPT subscriptions include features not available via the API (Deep Research, image generation, voice mode). Factor these into your comparison. The API is almost always cheaper for automated processing, while subscriptions are better for interactive human use.