OpenAI API: models and pricing

It depends on the model: from $0.030 to $150.00 per million input tokens. The cheapest is gpt-oss-20b ($0.030 / $0.130); the newest is GPT-6.1 Terra ($1.00 / $5.00). Data as of 11/14/2026.

No: the OpenAI API is pay-as-you-go, billed per million tokens. The cheapest model, gpt-oss-20b, costs $0.030 per million input tokens and $0.130 per million output tokens, so light use costs cents. Any free credits depend on OpenAI's own offer.

Yes, on 50 of its 64 models. For example, GPT-6.1 Terra charges $0.050 per million cached input tokens instead of $1.00: 95% less. The cache applies when you resend the same beginning of a prompt, like a long system prompt or a document.

For 10,000 requests a month of 1,500 input tokens and 400 output tokens: gpt-oss-20b costs about $0.97 and GPT-6.1 Terra about $35.00. Each model's page has a calculator for your own numbers.

64 chat models in the catalog, as of 11/14/2026. Each one has its own sheet with pricing, context and capabilities.

From the public catalogs of OpenRouter, models.dev and LiteLLM, which publish each provider's list prices. They sync once a day and the last update date is shown above the comparison.

It is what the provider charges per million tokens you send (input) or the model generates (output). A token is roughly 4 characters. Output almost always costs more than input.