Free model comparison resource

Token-saving worksheet

Compare two available models on the same representative tasks. Define acceptable output before testing, shorten unnecessary context, and record retries and correction time. Choose the lowest-cost option that passes your quality requirements.

1. Write the task, quality threshold and minimum context needed.
2. Run the same tasks with each model; record accepted results and failures.
3. For API use, record input/output tokens and actual billed cost, including retries. Divide total cost by accepted results.
4. For chat subscriptions, record plan cost and usage limits separately. Chat usage is not an API token bill.
5. Compare review time, privacy requirements and quality before choosing. Retest when your tasks or available models change.

The worksheet contains no assumed prices. Use your provider’s current rates and your actual usage.