How our Claude Code cost estimates work
Updated October 13, 2026
The calculator turns your answers into what your usage would cost at API rates. Then it compares that number with how much each plan holds. Anthropic publishes the ratios between plans but not the limits themselves, so one number is our own estimate: Pro holds about $125 of usage a month. We calibrated it on a real Claude Code user's logs. Every price comes from Anthropic's pages, with a link.
- Pro holds, at API rates (our estimate)
- $125/mo
- Calibration user, 30 days at API rates
- $1,497
- Prices last checked
- October 13, 2026
Step 1: what your usage is worth at API rates
Each kind of day has a cost at Opus API rates. The middle one sits close to Anthropic's published average of $13 per developer per active day. The heaviest one comes from the measured user below.
| A normal day | Cost at Opus rates |
|---|---|
| Quick help | $5 |
| Building features | $15 |
| All day, agents running | $60 |
The model you pick scales that cost by its output price compared with Opus ($20 per million output tokens):
| Model | Factor |
|---|---|
| Sonnet | 0.5x |
| A mix | 0.75x |
| Opus | 1x |
| Fable | 2.5x |
Then: cost per day x model factor x days a week x 4.33 weeks a month. For a team, we multiply by the number of people.
Step 2: how much each plan holds
Anthropic says Max 5x gives five times Pro's usage and Max 20x gives twenty times. A Team Premium seat gives five times a Standard seat. A Standard seat gives “more than Pro”, with no number, so we treat it as equal to Pro. That keeps Team picks on the cautious side.
| Plan | Compared with Pro | Holds about |
|---|---|---|
| Pro | 1x | $125 a month |
| Max 5x | 5x | $625 a month |
| Max 20x | 20x | $2,500 a month |
| Team Standard seat | 1x | $125 a month |
| Team Premium seat | 5x | $625 a month |
Step 3: how full each plan would be
Your usage divided by what a plan holds gives the fill bar:
| Fill | What we show |
|---|---|
| Up to 70% | Room to spare |
| 70% to 100% | Tight on busy days |
| Over 100% | You would hit limits |
The best fit is the cheapest plan with room to spare. Limits reset every five hours and every week, so a plan that is full on average runs out on busy days. The 70% line leaves room for them.
Where the Pro number comes from
We measured one developer's Claude Code logs with ccusage: $1,497 at API rates over 30 days, 26 active days, a median of $52 a day, mostly on Opus. They were on Max 20x and almost never hit a limit.
With Pro at $125, Max 20x holds $2,500, and this user fills 60% of it. That is inside the “room to spare” range, which matches what they saw. It is one data point, so treat results as a guide. You can check your own number on the usage page.
How we keep prices current
Every price on this site lives in one file, and each one links to the Anthropic page it came from. We last checked them on October 13, 2026. Each page shows the date it was last updated.
aicodecost is an independent site. It is not made by or affiliated with Anthropic.
Common questions
- Why not use Anthropic's real limits?
- Anthropic does not publish them as token or dollar amounts. It publishes the ratios between plans, so we estimate one number (Pro) and derive the rest.
- How accurate is the estimate?
- It is a guide. Your real usage depends on how much context Claude Code reads, which model you use and how long your sessions run. Measuring your own usage with ccusage gives a better input.
- Does the calculator include usage credits?
- No. It shows which plan holds your usage. If even the biggest plan fills up, it says you would need usage credits on your heaviest days.