Your usage. A different bill.
See what your OpenAI API usage would cost on other models. Same token volume. Today’s standard rates.
Bring your usage CSV
Drop it here, or choose a file to get started.
One CSV, up to 10 MB and 50,000 rows
Prepare your usage file
Select your report
Choose your organization, projects and period in OpenAI Usage.
Export by model
Export Responses / Chat Completions usage, grouped by model.
Check token counts
Include input and output tokens. A spend-only invoice won’t work.
From usage to alternatives.
Keep the volume. Compare the rates.
- 01
Bring your token counts
Export a usage CSV with model names, input tokens and output tokens. Or start with our demo.
- 02
Check the import
Review matched models and excluded rows. Adjust the columns if your export uses different headings.
- 03
Compare and take it with you
Inspect estimates for each source model, filter by provider and download the comparison as CSV.
Before you compare.
Does my file leave my browser?
No. Your CSV is read and processed locally, without an account or API key. Reloading or resetting the tool clears the imported data.
Is this my actual bill or guaranteed savings?
Neither. We apply current standard text API rates to the same input and output token counts for each model. Models are compared separately, without cache discounts. Tools, taxes, infrastructure and non-text usage are excluded.
Request-length tariffs can produce a range. Model quality, tokenization and output length can differ, so test alternatives on your workload before switching.
What if my CSV uses different columns?
Match your model, input and output columns after choosing the file. We show unmatched models and excluded rows for review. One CSV can contain up to 50,000 rows and must be no larger than 10 MB.
No usage file yet?
Plan your app’s monthly API budget from its expected workload.