01
Are you blocked now?
A refill can help immediately, even when its net gain over several days looks modest. The timing of important work matters.
Make your reset count.
One week's quota equals 100%. Daily use is evenly spread over each 24-hour block. Editing daily use applies that demand to every day in your current plan.
10% left · Refill in 24h · Need 70% / 24h
Codex
Restart clock
Claude
Keep date*
*Claude: keeping the date is an assumption.
Extra usage · % of one week’s quota
48-hour example · No reset: 80 / 140 units
Your usage plan
Extend the plan
Previously entered days are restored. This value fills only unentered days. Nothing changes until you confirm.
Planned demand / 24-hour block
One week's quota = 100. Demand can exceed 100; blank is unknown, not zero. Each block is evenly spread over 24 hours.
Same demand. Three paths.
WHY THE RESULTS DIFFER
Restarting serves 60 more units in the first 24 hours—but 40 fewer in the next. Net gain: 20.
Restart clock
of one week's quota, extra
100 units served
Keep the date
of one week's quota, extra
140 units served
48-hour comparison · One week's quota = 100 units.
| Block | Demand | No reset | Restart | Keep date |
|---|
Unused ending quota is shown separately, not counted as value. Unmet demand means “not served in the planned time,” not “impossible forever.” A refill exactly at the right endpoint is not included.
This is a weekly-quota simulation, not live account data. Unmet demand is not carried forward. This model ignores five-hour limits and other gates. It does not estimate task duration or guarantee real account availability.
01 / FOLLOW THE CLOCK
The clock is part
of the reset.
A reset changes more than a meter. One mechanism starts a new refill schedule. The other preserves the old date. The same top-up can meet different amounts of your planned demand.
Start with the no-reset path. Then compare all three paths over the same window, with the same demand—not three different periods chosen to flatter each result.
Scheduled refills, from the same starting point
EXAMPLE · NEXT 168 HOURS
Shaded area: the 48-hour comparison. Dots show refills, not extra work completed. Later refills are context, not part of the current result.
02 / THINK ABOUT YOUR WORK
Three questions before you reset.
03 / CHECK THE EVIDENCE
Rules, not assumptions
disguised as facts.
This calculator compares two explicit mechanisms. It does not claim every Codex or Claude reset follows one permanent rule.
These limited product facts are backed by official documentation. The account, offer and date still matter.
The cards illustrate two reset scenarios at equal weekly capacity. The Codex scenario models a paid instant reset with your first request made immediately; its new weekly period starts with that first request. The Claude scenario assumes an extra reset preserves the old refill date. That assumption is not established for every Claude reset by the regular Max-plan documentation. Check the terms of your account or offer.
Codex / banked resetOFFICIAL SOURCE · LIMITED SCOPE
A full banked reset refreshes five-hour and weekly windows and changes the weekly reset date.
Exact anchors, eligibility, scope and expiry must follow the account or offer; this is not a universal 168-hours-from-click rule.
Codex / paid instant resetOFFICIAL SOURCE · LIMITED SCOPE
The reset applies immediately, but the new weekly period starts at the first Work or Codex request afterward; the next reset is seven days after that request.
Do not generalize this anchor to every banked, hard or global event.
Claude Max / regular weekOFFICIAL SOURCE · LIMITED SCOPE
The regular weekly limit refreshes at an account-assigned time; other usage windows and restrictions also apply.
This regular-cycle description does not establish the behavior of every special Claude reset.
Rule notes reviewed 17 September 2026. A source supporting one field does not verify every setting in a model. Confirm the timing shown in your own account.
04 / THE IDEA BEHIND THE TOOL
More than
a fuller meter.
ILLUSTRATION / THE DEFAULT TWO-DAY SPRINT
Suppose your quota is 100 units and only 10 remain. A reset restores 90 units now. But “restored now” and “extra demand served” answer different questions.
With 70 units of demand in each 24-hour block, no reset serves 80 in total. Restarting serves 100. Keeping the date serves 140. The meaningful gains are the differences from 80—not the visual jump to a full meter.
These answers belong to this 48-hour plan. Change the demand, the original refill time or the observation window, and the outcome can change. In a longer plan, a later refill may align with a new demand peak.
A net total is not the complete value of your work. Deadlines, short-window caps, delayed tasks and model capabilities can all matter. Use this tool to understand the calendar effect, not as a promise about your account.
05 / SMALL PRINT, IN PLAIN WORDS
Before you act
on the number.
Does +20% mean 20% more productivity?
No. It means extra served demand equal to 20% of one normalized weekly quota. It is not a measure of productivity, quality, task success or relative improvement over the no-reset total.
Can the result be negative?
In the restart mechanism, yes. Moving a near-term refill can reduce the demand served inside the selected window. The calculator keeps that negative result visible.
Why not count unused quota as value?
A balance may help later, but it is not demand already served during this window. Ending balances are shown separately in the detailed comparison.
Are these live account numbers?
No. These are examples or values you enter. This page does not connect to your account, redeem resets, change your plan or save your inputs. Reloading restores the default example.
What does the model leave out?
Five-hour limits, model-specific restrictions, real task duration, priorities, changing capacity, future promotions and delayed-work queues are outside this model. Each entered day's demand is evenly distributed over its 24-hour block.
Is this a Codex-versus-Claude comparison?
The brand cards illustrate reset mechanisms at equal normalized capacity, not capabilities or plan allowances. Codex models a paid reset with immediate first use; Claude models an assumed keep-date reset. Actual reset terms depend on your account or offer.