Start on Claude Pro at $20 a month. If you keep hitting your limits, step up to Max 5x at $100. Go to Max 20x at $200 only if Max 5x runs out as well. Claude Code comes with all three. The free version of Claude does not include it.
The message you hit tells you which problem you have. “You’ve hit your session limit” means the five-hour window is spent, and Claude Code can wait for the reset and carry on by itself. “You’ve hit your weekly limit” can mean days of waiting. That second one is worth paying to avoid.
Anthropic says there is no fixed message count on its plans, so any “hours of Claude Code” figure, ours included, is an estimate. What Anthropic does publish is the price, a multiple for each five-hour session, and which models you get. This guide checks all of that against Anthropic’s own pages on October 26, 2026, shows what each plan means for Claude Code, and ends with a way to choose from your own numbers.
- Prices: Pro is $20 a month, or $200 billed up front for a year. Max 5x is $100 a month and Max 20x is $200 a month. Max has no yearly option.
- What Max adds: 5x or 20x Pro’s usage in each five-hour session, Fable models up to 50% of your weekly limit, and monthly API credits that cannot pay for interactive Claude Code.
- Claude Code is on all three: the same tool and the same default model, Opus 5.5. What changes is how long it runs before a limit stops it.
- Pick by the limit you hit: a five-hour stop clears in hours. A weekly limit that runs out midweek is what justifies the next plan up, or usage credits.
- Measure first:
/usagein Claude Code and Settings > Usage on claude.ai show both limits, and Context Board keeps them in sight on a board.
Claude Pro vs Max at a glance
Prices are in US dollars before tax. Everything in the table comes from Anthropic’s pricing page and its help center, read on October 26, 2026.
| Pro | Max 5x | Max 20x | |
|---|---|---|---|
| Price per month | $20 | $100 | $200 |
| Billed yearly | $200 up front, shown as $17 a month | Monthly only | Monthly only |
| Usage per five-hour session | The baseline | 5x Pro | 20x Pro |
| Weekly limit | Yes, across all models | Yes, across all models | Yes, across all models |
| Opus, Sonnet and Haiku | Included | Included | Included |
| Fable 5 and 5.1 | Usage credits only | Up to 50% of your weekly limits | Up to 50% of your weekly limits |
| Claude Code | Included | Included | Included |
| Usage credits past a limit | Yes | Yes | Yes |
| Monthly API credits | Not included | $100 | $200 |
| Anthropic’s description | For everyday work | Frequent users who work with Claude on a variety of tasks | Daily users who collaborate often with Claude for most tasks |
| Source | Pro plan article | Max plan article | Max plan article |
Do the sums on Anthropic’s own multiples. Call Pro’s session allowance one unit. Pro is $20 for one unit. Max 5x is $100 for five, which is still $20 a unit. Max 20x is $200 for twenty, which is $10 a unit, half the Pro rate. That holds for the five-hour session only. The weekly limit comes up below.
Claude Pro and Max prices
Pro is $20 a month. Pay for the year up front and it is $200, which Anthropic’s pricing page shows as $17 a month. Against twelve payments of $20, that saves $40.
Max 5x is $100 a month and Max 20x is $200 a month, per Anthropic’s Max plan article. Max is monthly only, so a year of Max 5x is $1,200 and a year of Max 20x is $2,400. The pricing page itself shows Max as “From $100” and says prices exclude tax.
These are US web prices. Anthropic says mobile app prices can differ and other countries see local currency. Past the yearly Pro price, there are no standing discounts on Pro or Max, though Anthropic says it occasionally runs limited-time promotions. Team and Enterprise plans are priced per seat, a separate decision this guide skips.
Upgrades take effect right away, and unused time on your old plan is credited toward the new one. If you pay for Pro by the year, that credit carries over to Max as long as the billing address matches. Downgrades take effect at the end of your billing period. You can cancel any time, but payments are generally non-refundable, and Anthropic says to cancel at least 24 hours before renewal to avoid the next charge. So testing Max costs one month of Max, not a year, and you can drop back to Pro when the period is up.
Max 5x vs Max 20x: how much more usage you get
Anthropic describes both Max tiers as multiples of Pro. Its Max plan article says Max 5x includes “five times the Pro plan’s per-session usage allowance” and Max 20x includes 20 times. A session is the five-hour window.
The multiple is attached to the session. Every paid plan also has a weekly limit that applies across all models, and the pages we read give no multiple for it. Max 20x is 20 times Pro for each five-hour session, but nothing we read says it is 20 times Pro for the week. If the weekly limit is what stops you, treat 20x as a session number and watch the weekly bar for a week on the new plan before you count on more.
The pool is shared. Anthropic says your activity across Claude on web, desktop and mobile, and in Claude Code, all draws from the same pool, and Claude Code in VS Code, Cursor or JetBrains counts too. A long afternoon in chat leaves less for coding. We cover how the session and weekly limits behave, and when they reset, in our guide to Claude’s usage limits.
Max’s card on the pricing page also lists higher output limits, early access to advanced features and priority access at high-traffic times, with no numbers for any of them. Anthropic adds that it may apply other limits, such as weekly and monthly caps or limits on a model or feature, at its discretion.
Limits have moved this year, so a number in an older article may be stale. On May 6, Anthropic doubled Claude Code’s five-hour rate limits on Pro and Max and removed the peak-hours reduction for Claude Code. On September 22, alongside Opus 5.5, it said it was increasing five-hour limits on Pro, Max, Team and seat-based Enterprise plans again, with no new numbers.
Which models Pro and Max include
Both plans include Opus, Sonnet and Haiku. In Claude Code the default model is Opus 5.5 on Pro, Max, Team and Enterprise, per the Claude Code model docs. That starts with Claude Code v2.1.280, released September 22. Earlier versions defaulted to Sonnet 5 on Pro.
Opus costs more of your limit than Sonnet. Anthropic’s guide to models and usage in Claude Code says Opus “uses meaningfully more of your quota,” calls Sonnet the right choice for the large majority of coding work, and offers /model opusplan, which plans with Opus and runs the edits on Sonnet. If Pro runs out too soon, that switch costs nothing.
Fable is where the plans split. On Max, Fable 5 and 5.1 are part of the plan, up to 50% of your weekly usage limits, per Anthropic’s Fable article. They draw from the same weekly limit and use it faster than other models. On Pro they are outside your limits and run on pay-as-you-go usage credits from the first message, and an interactive Claude Code session asks before it bills one.
You pick Fable with /model fable. It is not the default on any plan. The 50% caps how much of your week Fable can take. It does not add to your week, since Anthropic says you can never use more than your weekly limit.
Does Fable justify Max? Anthropic’s Opus 5.5 announcement says that model “performs at the level of Claude Fable 5.1 on most work.” The Claude Code docs call the Fable models the most capable in Claude Code, suited to tasks larger than a single sitting. In our reading, Fable alone is a thin reason to buy Max unless those long jobs are your work.
What Claude Code gets on each plan
Claude Code is part of Pro and Max. Sign in with your Claude login and it runs against the plan, with no separate bill. The terminal, VS Code, Cursor and JetBrains all count against the same limits as Claude chat, per Anthropic’s Claude Code article. It is the same tool on both plans, and several features that the docs tie to a plan, such as computer use and the three free /ultrareview runs, name Pro and Max together.
What differs between the plans:
- Room and Fable: the session multiples and the Fable access above.
- The wrap-up: if the five-hour limit lands mid-response, Claude Code may keep going for a moment to finish the step and show “Usage limit reached · wrapping up”. That usage counts toward your weekly limit. On Pro it applies up to once per weekly usage period. On Max it can apply each time you reach a five-hour limit. Anthropic is rolling it out gradually.
What Max does not change, and the last one is easy to miss:
- The wait: from v2.1.234, Claude Code can wait out a limit in the open session and continue the task by itself. It does not start that wait on its own when the reset is more than 24 hours away, which a weekly limit can be.
- Fast mode: it bills to usage credits on every subscription plan, Max included, per the fast mode docs.
- Monthly API credits: Max’s $100 or $200 cannot be spent in interactive Claude Code, or as usage credits. Anthropic’s article says they cover the Claude API, the Console playground, Managed Agents and the Agent SDK, and
claude -pwhen you run it with an API key from the linked Console organization. Do not count them as Claude Code room.
What to do when you hit a limit
You can wait, spend or move up.
Wait: the five-hour limit resets every five hours, and the weekly limit resets at a fixed time assigned to your account.
Usage credits: they are off until you turn them on in Settings > Usage, or run /usage-credits in Claude Code. Once they are on and you use up your five-hour limit, your weekly limit or, on Max, your Fable limit, requests switch to pay-as-you-go at standard API rates and keep working, until you reach your monthly spend limit or your balance runs out. You set the cap and prepay with Add funds or auto-reload, and the charge sits outside your subscription. Some Anthropic pages still call this extra usage, and the help center article has the steps.
Bundles: with usage credits on, you can pre-buy credits at a discount. A $50 bundle costs $45, a $250 bundle costs $200 and a $1,000 bundle costs $700, which is 10%, 20% and 30% off. Individuals can buy up to $2,000 worth of discounted bundles a month. Bundle credits only draw down after you pass your plan’s limits, per Anthropic’s bundles article.
The API: a Console account bills per token with no plan limits, which Anthropic suggests for intensive coding sprints. Watch for one trap. If ANTHROPIC_API_KEY is set on your machine, Claude Code uses the key instead of your subscription and bills API rates.
Move up: Anthropic’s own advice for Pro is to consider Max 5x if you consistently hit limits and need more capacity for larger repositories. For Max 5x it is Max 20x, on the same test.
For scale, Anthropic’s cost docs put the average Claude Code bill across enterprise deployments at about $13 per developer per active day, or $150 to $250 a month, at API rates. That is the neighborhood of Max 5x to Max 20x, so price a plan before you let credits carry a whole month.
Our rule of thumb on Pro uses the price gap. Max 5x costs $80 more than Pro. If your usage credits regularly pass $80 a month, price Max 5x. If they are $20 once in a while, stay on Pro and leave credits on as a safety net, with a spend limit set.
How to check your usage before you pay
Start here.
- In Claude Code, type
/usage. On Pro and Max it shows plan usage bars, activity stats and a breakdown of what used them, with a recommendation for anything behind 10% or more of your recent usage. Pressdorwto switch between the last 24 hours and the last 7 days. From v2.1.305, pressfto apply the fixes it can. - On the web, open Settings > Usage. At claude.ai/settings/usage, the plan usage limits section shows how much of your five-hour session you have used and how long it has left, when your weekly limit resets, and your usage credits.
Our guide to checking Claude Code usage covers both in more detail, plus the status line and what ccusage counts.
Two checks before you blame the plan. Run /status. A Login method row such as “Claude Max account” means Claude Code is using your subscription, per the Claude Code docs. And look at the model. Opus is the default, and routine work on Sonnet stretches a plan further.
Then try the free fixes. Anthropic lists five habits that stretch usage: clear between tasks with /clear, match the model to the job, point at files instead of pasting them, keep CLAUDE.md lean, and ask for a plan before big changes. Scheduled tasks and loops draw down the same usage as a session you type into, and /usage lists the heaviest ones. If you still run out with those in place, the limit is real.
Keep a tally for a week or two. Each time a limit message appears, write down which one it was and what day it landed. Heavy parallel work matters here. Anthropic’s docs say a single burst of activity, such as a large workflow fanout, can use up the weekly allowance before the session window resets.
Which plan to pick
Max vs Pro comes down to which limit stops you. Picture Thursday. Your week resets Monday at 9:00 AM and /usage says 6% left. No amount of waiting until dinner fixes that. A five-hour stop at 3:00 PM is a different animal, because the window comes back within hours.
The hours below are our rule of thumb, not Anthropic’s numbers. Hours are a rough guide, so the middle column matters more.
| Your week with Claude Code | What you see | Pick |
|---|---|---|
| Up to about 5 hours | A limit message is rare | Pro. Pay yearly ($200) if you plan to stay. |
| About 5 to 15 hours | The five-hour limit stops you now and then, and the week never runs out | Pro. Use Sonnet for routine work and let Claude Code wait out the reset. |
| About 15 to 30 hours | The five-hour limit stops you several times a week, or the weekly limit runs out | Max 5x. Try it for a month. |
| 30 hours or more, or several sessions at once | Max 5x’s weekly limit runs out before it resets | Max 20x. Watch the weekly bar for the first week. |
A few cases the table does not cover:
- Only Fable stops you on Max: move that work to Opus 5.5, or keep Fable going on usage credits.
- Your credits on Pro pass $80 a month: price Max 5x, as above.
- Max feels like too much: downgrade. It takes effect at the end of your billing period.
- You are weighing a second tool instead of a bigger plan: Claude Code vs Codex compares prices and limits for both.
If you are unsure, stay on Pro for two weeks and note which limit stops you. Then pay for one month of Max 5x and compare.
Keep both limits in sight on a board
Everything above waits for you to look. A board can keep the numbers in view instead.
Context Board is ours, so weigh the bias. It is a live board your agents keep current, and its AI usage widget, which you add from the gallery, shows what is left of your Claude Code five-hour window and of your week. Each is a ring that empties as you use it. It is green at 50% left or more, amber from 20% and red under 20%, with when it resets beside it, like “resets Mon 9:00 AM”. If your plan has a model’s own weekly limit, that shows too, and Codex sits in the same widget. The numbers are as each app last checked.
- Claude Code · docsasks to run the test suite3m
Week ring red by Thursday, two weeks running? Try Max 5x for a month.
For this decision, the weekly ring answers the question. If it is still green on Thursday, the weekly limit is not what stops you. If it is red by Tuesday, it is, and that is the limit you would be paying to avoid.
A small program on your computer reads the numbers. You set it up once with one line from Connect, and the Agents template has AI usage on it already. Only names, states and percentages leave your computer, never what you and Claude Code wrote. Our Claude Code dashboard guide compares this with Grafana and ccusage. Context Board is free for 3 days, then $12 a month.