AI API Request Cost Calculator
Paste your prompt, pick a model and instantly see token counts and cost — per request and per month.
Detected: 0 words ≈ 0 tokens (approx.)
We estimate tokens as 1 word ≈ 0.75 tokens. Actual usage depends on tokenizer and language.
Supported models & prices
Approximate prices per 1K tokens from public provider documentation. Always verify before production.
| Provider | Model | Type | Input / 1K | Output / 1K | Context | Notes |
|---|
How AI API request costs work
Every AI API call has two costs: input tokens (your prompt and system message) and output tokens (the model's reply). Providers charge these at different rates — output is typically 3–5× more expensive than input.
Token count is the key variable: 1 word ≈ 0.75 tokens (varies by language and model). A 200-word system prompt is roughly 150 tokens; a 500-word answer is ~375 tokens. This calculator estimates both and shows the total cost per request, plus a monthly projection if you specify request volume.
For multi-step agent workflows — where each step re-sends the full conversation history — the costs multiply quickly. See our AI Agent API cost calculator to model those scenarios.
Frequently asked questions
How accurate are these cost estimates?
We use official per-token prices from each provider. Token counts are estimated at roughly 0.75 tokens per word — actual tokenization varies by model and language, but results are within the right order of magnitude for planning purposes.
What is the difference between input and output token pricing?
Every provider charges separate rates for input tokens (your prompt) and output tokens (the model's reply). Output tokens are typically 3–5× more expensive. This calculator shows both costs separately so you can see the true breakdown and optimize accordingly.
Do you store my prompts or data?
No. All calculations run entirely in your browser. We collect only anonymous page-view analytics. Your prompts and files never leave your device.
How do I calculate monthly AI API costs?
Enter your expected number of requests per month in the "Requests per month" field. The calculator multiplies the per-request cost by that volume to give you a monthly estimate. For agent workflows with many steps, use the Agent cost calculator instead, as context accumulation dramatically changes the cost profile.
Which AI model is cheapest for my use case?
For simple chat or Q&A, GPT-5 Nano, Gemini 2.5 Flash-Lite or Mistral Small 3 are the cheapest options. For high-quality production outputs, Gemini 2.5 Flash or Claude Haiku 4.5 offer excellent price-to-quality ratio. Use the model comparison tool to run side-by-side cost estimates on your specific prompt.