#1 in LLM APIs & inferenceUsage-based
OpenAI API
The default API for GPT models and OpenAI tools.
The OpenAI API, on platform.openai.com, is the developer platform for OpenAI models, including text, image, and tool-use APIs. You pay per token or per call. It is a different product from ChatGPT.
Why this rank
Most builder tutorials, SDKs, and examples assume this API. It is the first platform to wire up unless you already need a specific other model.
Best for
Shipping a feature on OpenAI models.
Pricing
Pay per token, with prices published per model on the OpenAI pricing page. Set a usage limit on the project before you put a key in an app. ChatGPT subscriptions do not pay this bill.
Pricing checked November 7, 2026. Confirm it on the official site.
Before you commit
- A key in a client-side app will be stolen. Call the API from a server.
- Spend follows traffic. A usage cap is part of the setup, not an extra.
Also look at
- Claude Platform — Anthropic's API for Claude models.
- OpenRouter — One API key for many models and providers.
- Gemini API — Google's Gemini models, with a free tier in AI Studio.