Guide · Claude plans

Claude Max vs Pro: which plan for Claude Code?

Five times, twenty times, but of what? Here’s what each plan actually includes, and a simple way to decide based on how often limits stop you.

By Thomas Kanze

Last checked October 9, 2026. Limits change; the sources linked below are the final word.

See how Quotchi works

Quotchi comparing a small backpack with a much bigger one

The direct answer

Stay on Pro until limits cost you real work every week. Anthropic’s own advice is to move from Pro to Max 5x if you consistently hit limits, and from Max 5x to Max 20x if you still do. Max 5x ($100 a month) gives five times Pro’s per-session allowance; Max 20x ($200) gives twenty times. Both still have weekly limits, and Anthropic doesn’t describe those as a 5x or 20x multiple. If you only hit a limit now and then, usage credits or simply waiting is usually cheaper than upgrading.

Pro vs Max 5x vs Max 20x

What Anthropic documents for each individual plan (October 9, 2026)
	Pro	Max 5x	Max 20x
Listed price	$20/month, or $17/month billed yearly ($200 up front)	$100/month	$200/month
Per-session allowance	Baseline	5× Pro’s	20× Pro’s
Session limit	Resets every five hours	Resets every five hours	Resets every five hours
Weekly limit	Yes	Yes, across all models, fixed weekly time	Yes, across all models, fixed weekly time
Fable weekly limit	Not listed	Separate weekly limit	Separate weekly limit
Claude Code	Included	Included	Included
Usage credits	Available	Available	Available
Anthropic’s description	“More usage” than Free	“Ideal for frequent users who work with Claude on a variety of tasks”	“Ideal for daily users who collaborate often with Claude for most tasks”

Max prices are for web subscriptions, per Anthropic’s help center. Limits are shared across claude.ai, Claude Desktop and Claude Code on every plan, so heavy chat use also eats into your coding allowance.

The multipliers Anthropic publishes are per session. That’s the honest way to read “5x” and “20x”.

An honest rule of thumb

No one outside Anthropic can tell you how many prompts a plan buys, because there’s no fixed number. So base the choice on what you can observe: how often limits actually stop you. This is our rule of thumb, not Anthropic’s:

How often limits stop you, and what we’d try
What you see in a typical week	What we’d do
Rarely or never hit a limit	Pro. Annual billing if you’re sure you’ll stay.
Session limit once or twice, at awkward moments	Pro, plus habits: `/clear` between tasks, auto-continue on, heavy work right after a reset.
Session limit 3+ times, and it blocks real work	Try Max 5x for a month and see if the interruptions stop.
Weekly limit before the week ends	Max 5x, or usage credits for the last day or two if it’s rare.
On Max 5x and still hitting limits most days	Max 20x, as Anthropic suggests for consistent hits on 5x.
Running parallel agents or agent teams all day	Probably Max 20x, and check the `/usage` breakdown for subagent share first.

Count honestly. One afternoon of waiting for a reset can be fine if you had other work to do. Three blocked afternoons a week usually isn’t.

Worked example

Dana is on Pro. Over two weeks she writes down every limit message: five session limits, all mid-refactor, and no weekly limit. She turns on usage credits with a monthly cap and sees what the overflow costs her. In the first month her credits spend comes close to the $80 gap between Pro and Max 5x, so she switches to Max 5x. Her colleague hits the session limit about once every two weeks. For him, credits cost almost nothing and Pro stays the better deal.

The $80 is just the difference between the listed prices. Credits are billed at standard API rates, so your own spend is the number to compare.

Measure one normal week first

Gut feeling overstates limits, because the one time you got stopped mid-deploy is the one you remember. A week of notes gives you a better answer than any comparison table:

Write down every limit message. Note whether it said session limit, weekly limit, or a model name, and what you were doing. A model-specific limit is solved with /model, not a bigger plan.
Check /usage at the end of each day. The plan bars show where the session and weekly windows stand. On paid plans the breakdown shows what used the allowance over the last 24 hours or 7 days (press d or w).
Look at what was heavy. If one MCP server, a /loop or a subagent setup accounts for a big share, fix that first. It’s cheaper than any plan change.
Note how much chat you did. claude.ai and Claude Desktop draw on the same limits. A research-heavy week in chat shows up in your coding allowance.
Decide with the table above. If the week was unusual (a launch, a migration), measure another one.

The status line can keep the five-hour and weekly percentages in front of you while you work, which makes the notes easier. Here’s the five-minute setup.

Questions that settle it

Is it the session limit or the weekly one? Session-only pain often goes away with habits and auto-continue. Weekly pain usually needs more allowance.
Do you have something else to do while you wait? If a reset is a coffee break, Pro may be fine. If it’s a dead stop, it isn’t.
Is it every week or one bad week? Credits are made for the occasional bad week.
Are you running agents in parallel? Every subagent and teammate sends its own requests against the same limits.

Is Max “worth it”? Doing the math

Two numbers help:

Your usage-credits spend. If you already pay for credits on Pro, compare a typical month’s spend with the price difference to the next plan. When credits regularly approach it, the upgrade is likely better value. When they don’t, stay.
The API-equivalent value of your usage. Claude Code’s /usage Session block shows an estimated dollar figure at list price for API users. On a subscription it isn’t your bill, but it shows how much work you’re doing. Our Claude API value calculator turns token counts into that kind of estimate so you can sanity-check a plan.

Don’t forget the soft costs. A blocked session at 4pm on a deadline costs more than its token price. If limits regularly interrupt deep work, that’s a fair reason to upgrade even when the math is close.

Try these before you upgrade

Clear more. Each request resends the whole conversation. /clear between unrelated tasks costs nothing.
Use the cheaper model for routine work. Anthropic’s cost docs suggest Sonnet for most coding and Opus for complex reasoning, and Haiku for simple subagent tasks.
Avoid cache misses. On a subscription the main conversation’s cache lasts an hour. Long breaks mid-session reprocess everything.
Check the breakdown. /usage on paid plans attributes recent usage to skills, subagents, plugins and MCP servers, and flags long context or cache misses.
Keep auto-continue on. Waiting out a session limit costs nothing if Claude Code resumes on its own.
Watch your chat use. Long claude.ai conversations use the same allowance.

And keep an eye on how close you are. Quotchi shows your five-hour and weekly Claude Code usage in the Mac menu bar, and Quotchi Pro adds a weekly plan coach that looks at your own readings over time.

Not sure how often you really hit limits? Quotchi shows Claude Code’s session and weekly usage in the menu bar, with alerts you set. Free, local-first, no Quotchi account.

See how Quotchi works

What about Team?

Team plans work differently. Each member draws on a per-seat allowance with a rolling five-hour window and a weekly window, shared with Claude chat and Cowork. Anthropic’s pricing page lists Standard seats at $25 per seat monthly ($20 billed yearly) and Premium seats at $125 ($100 billed yearly), with Premium described as 5x more usage than Standard. Admins can turn on usage credits with spend limits at the organization, group or member level.

Switching plans

You can move between Pro and Max from claude.ai. Inside Claude Code, signing back in after /upgrade while a usage-limit wait is running makes Claude Code check right away whether usage is available again. If you upgrade in a browser on your own, it doesn’t re-check automatically; send a prompt to continue.

If you’re undecided between plans, the builder quiz takes under a minute and gives a plan suggestion based on how you work.

Common mistakes and myths

“Max 20x means 20 times the weekly limit.”: Anthropic says 20 times Pro’s per-session allowance. It doesn’t describe the weekly limit as a multiple.
“Max removes limits.”: No. Max has session limits, weekly limits and a separate Fable weekly limit.
“Max’s API credits pay for Claude Code overflow.”: Anthropic says monthly API credits included with Max can’t be used in Claude Code or the Claude apps.
“There’s a published prompt count per plan.”: There isn’t. Usage depends on conversation length and complexity, features, model and effort level.

What changed recently

As of October 9, 2026, Anthropic’s plan article lists a separate weekly limit for Fable on Max plans, and the pricing page lists Pro at $17 a month when billed yearly. Claude Code also gained automatic continue after a usage limit (v2.1.234), which makes waiting out a session limit on Pro less painful than it used to be.

Sources

Official pages we read on October 9, 2026. If anything here disagrees with them, they win.

FAQ

How much more usage does Max give than Pro?

Max 5x includes five times and Max 20x twenty times the Pro plan’s per-session usage allowance, according to Anthropic. Weekly limits still apply and aren’t described as multiples.

Is Max worth it for Claude Code?

If limits interrupt real work several times a week, probably. Anthropic suggests upgrading when you consistently hit limits. If it’s occasional, usage credits or waiting is usually cheaper.

Do I still get weekly limits on Max?

Yes. Max has a weekly limit across all models that resets at a fixed weekly time, plus a separate weekly limit for Fable.

How many messages do I get on Pro or Max?

Anthropic doesn’t publish a fixed number. It depends on conversation length and complexity, features, model and effort level.

Can I use Max’s API credits in Claude Code?

No. Anthropic says monthly API credits included with Max can’t be used in Claude Code or the Claude apps. Use usage credits instead.

What does Max cost?

Anthropic lists Max 5x at $100 a month and Max 20x at $200 a month for web subscriptions. Pro is $20 a month, or $17 a month billed yearly.

Should I pick Max 5x or Max 20x?

Start with Max 5x unless you already know 5x isn’t enough. Anthropic suggests moving to 20x if you consistently hit limits on 5x.

Does chatting on claude.ai use my Claude Code allowance?

Yes. Anthropic says limits are shared across Claude and Claude Code, so all activity in both counts.