Tokens × API list price
That is the amount on the card and the ranking. Not what you paid: what the same use would have cost through the API.
How it counts
| Item | Claude Code | Codex |
|---|---|---|
| Token types | Input, output, cache reads and cache writes | Uncached input, cached input and output |
| Prices | Anthropic API list prices | OpenAI API list prices |
Good to know
- Use in the Codex and Work apps on a computer counts. Plain Claude and ChatGPT chats, and Codex or Work used on a phone or on the web (chatgpt.com), leave no record on the computer and cannot be counted.
- Prices are the list prices of October 2026. A model without a public price is estimated at a similar model's rate. The exchange rate is Hana Bank's posted rate as Naver shows it.
- Results may be incomplete if usage records are missing or unavailable on your computer.
- Your conversation text and source code are not uploaded. See the privacy policy for what is collected and stored.
The ranking's amounts
The server works it out again at the same list prices, so the plans and exchange rate you change on your card do not reach the ranking.
One computer's record is never counted twice or put up again under another account. If you use several computers, sign in on each and they add up. What is stored and shown is in the privacy policy.
Whale tiers
Your place within the selected period and scope sets your image tier. At 100 participants: Challenger #1; Grandmaster #2–3; Master #4–6; Diamond #7–13; Emerald #14–25; Platinum #26–40; Gold #41–60; Silver #61–80; Bronze #81–95; Iron #96–100.
With fewer than 10 participants, each place gets the next tier starting from Challenger. At 10 or more, every tier gets at least one participant; remaining places follow the target proportions. These tiers reflect your usage ranking on TokenWhale.