LLM Token Calculator

Estimate token count and API cost from text length across different tokenizers (GPT-4, Claude, Llama).

Text length (characters)

Tokenizer

Input price per 1M tokens ($)

Output-to-input ratio

Number of requests

Tab to move between fields

Input tokens per request

1,250

Total API cost ($)

$0.02

Output tokens per request1,250

Total tokens per request2,500

Total tokens (all requests)2,500

Input cost ($)$0.00

Output cost ($)$0.01

128K context usage (%)1.95%

Tokenizer NameGPT-4 (cl100k_base)

How to Use This Calculator

Enter the Text Length in characters — paste your typical prompt and count characters, or use a rough estimate.
Select the Tokenizer matching your model (GPT-4 uses ~4 chars/token, Claude ~3.5 chars/token).
Set the Input Price per 1M Tokens from the provider's pricing page and the Output-to-Input Ratio (e.g., 0.3 for summarization, 2 for generation).
Enter the Number of Requests to estimate total batch cost.
Review Input Tokens, Output Tokens, Total Cost, and Context Usage % to optimize your prompt length and budget.

How the result changes with Text length (characters)

Text length (characters): Number of characters in your input text. Average English word is ~5 characters.
Tokenizer: Select the tokenizer
Input price per 1M tokens ($): Cost per 1 million input tokens. GPT-4o: $2.50, Claude Sonnet: $3, GPT-4 Turbo: $10.
Output-to-input ratio: Expected output tokens as a multiple of input tokens. Summarization ~0.3, Q&A ~1, generation ~2+.
Number of requests: Total API requests to estimate cost for.

Input tokens per request: Estimated input token count per request.
Output tokens per request: Estimated output token count based on the output ratio.
Total tokens per request: Input + output tokens consumed per request.
Total tokens (all requests): Grand total token consumption across all requests.
Input cost ($): Cost for input tokens across all requests.
Output cost ($): Cost for output tokens (typically 3x input price).
Total API cost ($): Combined input + output token cost.
128K context usage (%): Percentage of a 128K context window used per request.

Identify Input Parameters
4 parameters
Text length (characters) = 5000, Tokenizer = 0, Input price per 1M tokens ($) = 3, Output-to-input ratio = 1 = 5 input(s) provided
Calculate Input tokens per request
1250 = 1250
Calculate Total API cost
Total API cost = inputCost + outputCost
0.015 = $0.015
Calculate Output tokens per request
Output tokens per request = ceil(inputTokens * outputRatio)
1250 = 1250
Calculate Total tokens per request
Total tokens per request = inputTokens + outputTokens
2500 = 2500