AI Cost Calculator
Paste in a prompt (or just estimate its length) to see what it would actually cost to run across current OpenAI, Anthropic, and Google API pricing — as a single request and at monthly volume.
Pricing last verified 2026-08-07 directly from each provider's own pricing page — see sources below. Runs entirely in your browser; nothing you type is sent anywhere.
0 estimated input tokens
Estimated cost by model
| Model | Per request | Per month |
|---|
Sorted cheapest to most expensive at your current monthly volume. Excludes provider-side extras like prompt caching discounts, batch API discounts, or tool-use overhead.
Why the same prompt costs such different amounts
AI API pricing is set per model, not per provider — every major provider sells a range of tiers from a flagship reasoning model down to a lightweight, high-volume model, and the price gap between the top and bottom of that range is typically 20-25x. A task that a budget-tier model handles well can cost a small fraction of running the same prompt through a flagship model, which is why matching model choice to task complexity is one of the more effective ways to control real API costs at volume.
Sources
- OpenAI pricing: https://platform.openai.com/docs/pricing
- Anthropic pricing: https://docs.anthropic.com/en/docs/about-claude/pricing
- Google pricing: https://ai.google.dev/gemini-api/docs/pricing
Frequently asked questions
How accurate is the token estimate?
It's an approximation based on character count (roughly 4 characters per token in English), the same rule of thumb providers themselves publish. Real tokenization varies slightly by model and language, so treat the result as a close estimate, not an exact bill.
Why do prices differ so much between models?
Providers price each model tier based on its capability and compute cost — a flagship reasoning model can cost 20-25x more per token than that same provider's budget tier, which is why matching a task to an appropriately sized model matters for cost.
Does this tool send my text anywhere?
No — the entire calculation runs in your browser using JavaScript. Nothing you type is sent to a server, logged, or stored.