LimitPlanner
Menu

The usage forecast · checked 20 Oct 2026

Will your week fit your AI plan?

Plan your week and get the forecast: the day you'll likely run dry, which days to take off, the cheapest plan that fits and what the overflow costs. Honest low, likely and high ranges; every figure sourced.

  • 24 plans
  • 7 vendors
  • No sign-up, no cookies
weekly cap
1

Your plan

Claude Max 5x $100/mo billed monthly

About $550 API-$ a week inferred

High confidence Limits and sources

2

Your week

Start from a sample week

4 hours of interactive agentic coding, 15 quick questions and 2 PR reviews every weekday.

How many of each task you plan per day, Monday to Sunday
TaskMonTueWedThuFriSatSun
Agentic coding, 1 hour, one interactive agent$2–$25 each
Quick chat Q&A$0.01–$0.05 each
Code review of a pull request$0.20–$1.50 each
3

The forecast

ForecastClaude Max 5xHigh confidence

Clear skies all week.

Fits under the weekly cap even in the high case.

  1. Mon (Monday) ClearUnder the weekly cap $33.30$8.55–$103.75
  2. Tue (Tuesday) ClearUnder the weekly cap $33.30$8.55–$103.75
  3. Wed (Wednesday) ClearUnder the weekly cap $33.30$8.55–$103.75
  4. Thu (Thursday) ClearUnder the weekly cap $33.30$8.55–$103.75
  5. Fri (Friday) ClearUnder the weekly cap $33.30$8.55–$103.75
  6. Sat (Saturday) ClearNothing planned $0no API-$ work
  7. Sun (Sunday) ClearNothing planned $0no API-$ work
Cumulative API-$ through the week against the weekly cap of $550.
Weekly cap $550

Week total

Low
$42.75
Likely
$166.50
High
$518.75

API-$ equivalent against a weekly cap of $550 (inferred).

Off-days

No off-days needed. The week fits even in the high case.

5-hour sessions

2 windows × $78.60 (derived) = $157.20 a day.

Every day fits inside 2 windows in the likely case.

Overflow

No overflow, even in the high case.

Usage credits at standard API rates once you hit a limit.

Cheapest plan that fits

Claude Max 5x $100/mo

Cheapest plan whose likely week fits, including the 5-hour session check.

Fits even in the high case: Claude Max 5x $100/mo.

Compares plans with a dollar capacity (Claude and GitHub Copilot). Ranked by price only; affiliate links never change it.

Every plan we can compute, for this week

PlanPrice/moConfidenceFits (likely)Runs dry (likely, range)Overflow (likely)
GitHub Copilot Pro$10High confidenceNoMon (Mon to Mon)$129.74
ChatGPT Plus (Codex)$20Medium confidenceNo Codex rows–not published
Claude Pro$20High confidenceNoTue (Mon to never)$100.50
GitHub Copilot Pro+$39High confidenceNoMon (Mon to Wed)$117.03
Claude Max 5x$100High confidenceYes, even highnever–
GitHub Copilot Max$100High confidenceNoTue (Mon to never)$87.01
Claude Max 20x$200High confidenceYes, even highnever–

Sorted by price. Plans with only relative limits (X, Google, Cursor, Perplexity, ChatGPT Pro) can't be forecast from public data; calibrate one in the calculator.

How to read the forecast

Each day is a forecast card. The bar under it is your tank: the dot is the likely running total, the shaded band runs from the low to the high case, and the dashed red line is the weekly cap.

  • ClearUnder the weekly cap, even in the high case.
  • CloudyFits in the likely case; the high case goes over.
  • StormThe likely running total is over the weekly cap.
  • SquallThe day needs more 5-hour windows than you use.
  • Off-dayTake this day off (or move its work) and the week fits.
  • FogNo public number to forecast against; calibrate it.

Questions people ask

How sure are we, per vendor?

  • Claude (Anthropic) High confidenceOfficial multipliers plus two independent measurements of absolute capacity (she-llac credit table, alldonesites live meter). Pro and Max 5x weekly figures are inferred, Max 20x is measured.
  • GitHub Copilot High confidenceGitHub publishes hard numbers: monthly AI Credits in dollars (1 credit = $0.01).
  • ChatGPT / Codex (OpenAI) Medium confidenceOpenAI publishes message ranges per 5 hours for Codex on Plus. Pro tiers only get multipliers and an unpublished weekly allowance.
  • X Subscription (Grok + Cursor) Low confidenceMultipliers only (4x, 15x, 50x Plus). The size of the base weekly pool is not published.
  • Cursor (legacy plans) Low confidenceMonthly pools; only rough dollar guidance is published.
  • Google AI (Gemini) Low confidenceCompute-based multipliers only; hard counts exist only for Flow credits and NotebookLM.
  • Perplexity Low confidenceThe official limits page was blocked for our fetcher; counts are unverified.

Read the methodology for every assumption, or browse all plans with their sources.

Week doesn't fit?

Some one-job tasks (meeting notes, short copy, voice-overs) can move to a specialist tool outside your AI plan's quota. See the offload list.