API Pricing
OpenAI API Cost Calculator
Estimate your OpenAI API costs by model, token count, and daily request volume — see per-request, daily, monthly, and annual spend side by side. AI costs scale fast: a chatbot processing 10,000 queries/day on GPT-4o costs ~$750/month in API fees alone.
Choosing the right model is the biggest cost lever — GPT-4o mini costs 17× less than GPT-4o per token and handles most summarisation, classification, and extraction tasks at comparable quality. Use the presets to benchmark common use cases like chat, RAG queries, and code generation.
Advertisement
OpenAI API Pricing Explained
OpenAI charges separately for input tokens (your prompt) and output tokens (the model's response). One token ≈ 4 characters or ¾ of a word in English. A typical ChatGPT conversation turn uses 200–500 input tokens and 100–300 output tokens.
| Model | Input /1M | Output /1M | Best for |
|---|---|---|---|
| GPT-4o mini | $0.15 | $0.60 | High-volume, cost-sensitive apps |
| GPT-4o | $2.50 | $10.00 | Balanced quality & cost |
| o1-mini | $3.00 | $12.00 | Reasoning tasks, moderate cost |
| o1 | $15.00 | $60.00 | Complex reasoning, research |
| GPT-4 Turbo | $10.00 | $30.00 | Legacy high-quality tasks |
For most production applications, GPT-4o mini offers the best cost-to-quality ratio. Reserve GPT-4o for tasks requiring nuanced understanding, and o1 only for complex multi-step reasoning where accuracy is critical.
Frequently Asked Questions
How do I count tokens before making an API call?▼
Use the tiktoken library (Python: pip install tiktoken). For GPT-4o: encoding = tiktoken.encoding_for_model('gpt-4o'); tokens = len(encoding.encode(text)). Alternatively, the OpenAI Tokenizer Playground at platform.openai.com/tokenizer shows exact token counts interactively.
What is the context window and how does it affect cost?▼
The context window is the maximum total tokens (input + output) a model can process in one call. GPT-4o supports 128K tokens. Every token in your message history counts as input tokens — long conversations become expensive. Use context compression or summarisation for long-running agents.
Does OpenAI offer cheaper batch pricing?▼
Yes. The Batch API offers 50% discount on all models for asynchronous requests that complete within 24 hours. Ideal for classification, summarisation, or embedding tasks that don't need real-time responses. The same endpoint costs $1.25/1M input tokens for GPT-4o instead of $2.50.
How can I reduce my OpenAI API costs?▼
Key strategies: (1) Cache responses for identical inputs using a hash map or Redis. (2) Use GPT-4o mini for classification/extraction tasks. (3) Reduce system prompt length — it's charged on every call. (4) Use streaming to fail fast on bad outputs. (5) Batch API for offline processing. (6) Fine-tuning can produce a smaller model with similar quality at lower cost for specific tasks.
What is the difference between tokens and words?▼
Tokens and words are not the same. In English, 1 token ≈ 0.75 words. Common words are often a single token; rare words and non-English text may use 2–4 tokens per word. 'Hello world' = 2 tokens; 'Electroencephalography' = 5 tokens; Chinese/Japanese characters use 2–3 tokens each.