How many words is 1,000 tokens?

In English prose, 1,000 tokens is approximately 750 words (or 4,000 characters). Code and JSON typically use slightly more tokens per character due to punctuation and indentation.

Why are output tokens more expensive than input tokens?

Input tokens can be processed in parallel during the prefill phase, whereas output tokens must be generated one by one sequentially, holding GPU KV-cache memory throughout generation.

AI Token & Cost Calculator

Estimate tokens and API cost for AI models like Claude. Free, runs in your browser.

In plain words: AI models charge by "tokens" (roughly ¾ of a word). Paste text to see about how many tokens it uses and what it might cost. Counts are estimates and prices change often.

🔒 Runs in your browser. Nothing you type is sent anywhere ✓ No email · No sign-up · No watermark · Free

Prices as of October 2026 — always check the provider's pricing page. Figures are approximate list prices (USD per 1M tokens, standard tier, no caching/batch discounts) and can change at any time.

Input (Prompt) Tokens per Call Output (Completion) Tokens per Call API Calls per Day

Optional: Paste a Prompt to Measure Live Token Count ~0 tokens

Model Pricing Comparison (Per Call, Daily & Monthly @ 30 Days)

Your own model: input $ per 1M tokens Your own model: output $ per 1M tokens

AI Model	Input / 1M	Output / 1M	Cost per Call	Daily Cost	Monthly (30d)

🔗 Chain next workflow: Jump straight to a related tool

✉️ Paste ChatGPT Answers Cleanly into Email or Word { } JSON Formatter, Validator & Tree Explorer 💳 Stripe & PayPal Fee Calculator (UK)

How to Use AI Token & Cost Calculator

Enter your average input tokens, output tokens and API calls per day, or paste a sample prompt to get an estimated token count.
Compare the estimated cost per call, per day and per month for the listed models.
Add the input and output prices of your own model per million tokens to compare any other provider. Prices change often, so check each provider’s pricing page.

How LLM Token Pricing Works for SaaS & AI Workflows

Large Language Model APIs bill separately for Input (Prompt) Tokens and Output (Completion) Tokens per 1 million tokens (1M tokens ≈ 750,000 English words, or roughly 1 token ≈ 4 characters). The token count from a pasted prompt is an estimate based on its length, not an exact tokenizer, and listed prices are approximate standard-tier prices without caching or batch discounts. Output tokens typically cost 3x to 5x more than input tokens because generating tokens requires sequential autoregressive memory bandwidth on datacentre GPUs.