Understanding Token Calculations

Tokens are a fundamental unit used to measure the amount of text processed by our model. Here's how we approximate the usage and cost:

Token to Word Conversion:

• Tokens are calculated based on the text input to and output from the model.
• For instance, if you generate 32,000 tokens, this roughly translates to 24,000 words, as typically 3/4 of the tokens represent actual words.

Words to Pages Estimation:

• On average, each page contains about 500 words.
• Therefore, 24,000 words are equivalent to approximately 48 pages (since 24,000 words / 500 words per page = 48 pages).

Input and Output Token Relationship:

• For every input token, there is typically an equivalent output token. This means that the total number of tokens considered will be a combination of both input and output tokens.

Cost Calculation:

• The cost is computed based on the total tokens processed. This includes both the tokens for input and the tokens for output.
• Although this method provides a good baseline, please note that it is an approximation and might vary based on specific text characteristics and model behavior.

This approach gives you a straightforward way to estimate how much content you can process and the associated costs. Remember, this is a starting point, and you might need to adjust these estimates based on your actual usage.