tokens
What are tokens?
Tokens are the billing unit for LLM APIs. They are not the same as words or characters — models split text into sub-word chunks, so a short word might be one token while a long word might be two or three.
Providers charge separately for input tokens (your prompt, system message, and context) and output tokens (the model's reply). Output is usually priced higher per token than input.
how to estimate
Estimate cost in three steps.
- Open AI Token Cost Calculator and paste your prompt or document.
- Pick an expected output size — Short, Medium, Long, or a custom token count.
- Compare the cost table across GPT, Claude, Gemini, DeepSeek, Grok, and Perplexity models.
The tool uses a ~4 characters per token approximation. That is close enough for budgeting; exact counts depend on each model's tokenizer.
real bills
When estimates differ from your bill.
- Tokenizer variance — the ~4 chars/token rule can be off by 10–20% for some models and languages.
- Cached input — repeated context may bill at a lower cached rate on some providers.
- Tool calls and search — agent workflows add extra tokens and sometimes per-request fees.
- Batch and volume discounts — async batch jobs and enterprise tiers change the effective rate.
Rates in the calculator were last verified on 21 Jun 2026. Check the provider pricing links in the tool before you lock in a budget — prices change.
budgeting
Turn a per-run estimate into a monthly number.
Multiply the total cost for one run by how many times you expect to call the API per day, then by ~30 days:
Monthly estimate ≈ per-run total × daily runs × 30
If output length varies a lot, run the estimate twice — once with Short output and once with Long — and use the range for planning.
privacy
Browser-local only.
Your pasted text stays on your device. The calculator does not send prompts to Tiny Design Shop or any AI provider. Pricing is hard-coded and updated manually — there is no live pricing API call.