AI Token & Cost Estimator
Estimate token counts and API cost for GPT, Claude and other LLM prompts before you run them, free and calculated right in your browser.
Your usage
Token counts are approximate (no tokenizer). Pricing presets are manually maintained (June 2026) and editable — always confirm current pricing with the provider. USD. Nothing is uploaded.
Estimate
How to use this calculator
- Paste a typical prompt (or type the input token count directly) and set the expected output length in words.
- Pick a model preset — it fills in the input/output price, which you can edit to match your exact plan.
- Set how many runs per day and days per month you expect.
- Read the token estimate, context fit, and cost per run / day / month / year.
How it works
Tokens are approximated (roughly 4 characters per token, blended with a words-based estimate) — it's a planning figure, not the exact count a specific model's tokeniser produces. Output tokens are estimated from the words you expect back.
Cost per run = input tokens ÷ 1,000,000 × input price + output tokens ÷ 1,000,000 × output price. That scales by your runs/day, days/month and 365 days/year.
Pricing presets are manually maintained and approximate — they're editable so you can drop in your real per-million-token rates. Always confirm current pricing with the provider.
Worked example
A 2,000-token prompt, 300-word answer, 50×/day. At a Sonnet-class rate ($3 input / $15 output per 1M tokens), one run is roughly a fraction of a cent; the value of the tool is seeing the per-month and per-year totals as runs/day climb.
Common mistakes
- Treating the token count as exact — it's an approximation. Add a margin for safety.
- Forgetting output cost — output tokens are often priced several times higher than input tokens.
- Using a stale preset price — confirm the current rate with the provider and edit the field.
Frequently asked questions
Does this use AI or send my prompt anywhere?
No. It runs entirely in your browser using rule-based checks — your prompt is never uploaded or sent to any AI model. Still, avoid pasting passwords, API keys or private data into any tool.
Are the prices live?
No — they're manually maintained, approximate presets (last reviewed June 2026) that you can edit. There's no pricing API. Always verify the current rate with the provider.
How accurate is the token estimate?
It's a ballpark using a characters-and-words heuristic, typically within ~10–20% for English prose. For exact billing, check the provider's usage dashboard.
Related tools
- AI Context Window & Chunk Planner
- Prompt Score Calculator
- AI Model Selector Helper
- AI Business Idea Prompt Builder
- AI Structured Output / JSON Prompt Builder
- AI Data Analysis Prompt Builder
Explore more in AI Tools.
Cite this tool
Zerdly. (2026). AI Token & Cost Estimator. Retrieved from https://www.zerdly.com/tools/ai-token-cost-estimator/
Tip: Enter any known values to calculate the remaining results.
All calculations run in your browser. Your inputs are never saved or transmitted.