OpenAI API Cost Calculator 2026

OpenAI charges for API usage by the token, with input and output tokens priced separately. GPT-4o costs $2.50 per million input tokens and $10 per million output tokens. At typical usage of 1,000 input tokens and 500 output tokens per call, that is $0.0075 per call, or $7.50 per 1,000 calls. This calculator lets you model realistic monthly costs based on your actual usage patterns.

Enter the model, average tokens per input (your prompt plus any context), average tokens per output (the model response), and expected monthly call volume. The calculator gives you estimated monthly and annual costs, helping you budget API spend before scaling or choose between model tiers based on cost tradeoffs. Supports GPT-4o, GPT-4o mini, o1, o1-mini, o3, o3-mini, and GPT-4-turbo.

GPT-4o costs $2.50 per 1 million input tokens and $10.00 per 1 million output tokens as of June 2026. A typical short prompt of 500 tokens with a 300-token response costs around $0.0043 total.

The cheapest OpenAI model is GPT-4o mini at $0.15 per 1 million input tokens. For high-volume applications like chatbots or document processing pipelines, switching from GPT-4o to GPT-4o mini can reduce API costs by more than 90% with minimal quality loss on routine tasks.

Use the AI token counter to estimate your token counts before running cost projections. For cross-model comparisons, the LLM cost comparison tool shows costs across OpenAI, Claude, and Gemini side by side.

GPT-4o mini is the most affordable OpenAI model at $0.15 per 1 million input tokens and $0.60 per 1 million output tokens. It is suitable for summarization, classification, simple Q&A, and most writing tasks.

o1-mini and o3-mini cost $1.10 per 1 million input tokens and offer reasoning capabilities at a lower price than the full o1 and o3 models. Reserve o1 ($15.00 input) and o3 ($10.00 input) for tasks that require complex multi-step reasoning, math, or coding problems where accuracy is critical.

To compare these costs against Anthropic models, see the Claude API pricing calculator.

The OpenAI Batch API offers a 50% discount on input and output tokens for all eligible models. Requests are processed asynchronously and complete within 24 hours.

Batch API is ideal for workflows that do not need instant responses: generating product descriptions, analyzing documents overnight, or processing large datasets. You submit a JSONL file with your requests and retrieve results when processing is complete. The 50% discount applies to the same models as the standard API.