Claude Code Pricing 2026: Is It Free? Every Plan

Claude Code is not free: Anthropic leaves it out of the Free plan, and the cheapest subscription that includes it is Claude Pro at $20 a month, or $17 a month if you pay $200 for a year up front. Above that, Claude Code pricing depends on how much you use it: Max costs $100 or $200 a month, Team seats cost $20 to $125 each, and a Claude Console account skips the subscription and bills every token at API rates. Those figures come from Anthropic's Claude Code page and Claude pricing page.

We wrote this for people who don't code for a living but want Claude Code for reports, file clean-ups, small websites and automations. Prices and limits come from Anthropic's own pages as of October 2026, and each claim links to its source. You also get the two narrow ways to try Claude Code without paying, what the five-hour and weekly limits mean for your working week, and a calculator that turns your own usage bar into hours. For Claude's chat plans in general, start with our full Claude pricing guide.

Is Claude Code free? No, with two narrow exceptions

No. Claude Code requires a Pro, Max, Team, Enterprise or Claude Console account, or access through a cloud provider such as Amazon Bedrock, and Anthropic's setup documentation says "the free claude.ai plan does not include Claude Code access." The rule covers the terminal and code-editor versions, the Code tab in the Claude desktop app and Claude Code on the web, which needs a Pro, Max or Team plan or an eligible Enterprise seat. Click the Code tab on a Free account and the desktop app asks you to upgrade first.

Two official programs give you Claude Code at no cost, and both come with conditions:

• A seven-day guest pass.Eligible Pro and Max subscribers can share a limited number of passes, and each one gives "seven days of free access to Claude's Pro plan, which includes Claude Code, Cowork, and Claude Design," according to Anthropic's guest pass article. You must be new to paid Claude plans and enter payment details at sign-up, and the trial turns into a paid Pro plan unless you cancel before the seven days end.
• Claude for Open Source. Open-source maintainers and active contributors can apply for "6 months of free Claude Max 20x". The page's "Who should apply" list sets a high bar: 500 or more dependent repositories, 200,000 or more combined monthly downloads, or 100 or more pull requests merged into repos you don't own in the last 12 months are three of its examples. Most readers of this blog won't fit that profile.

Two other things look free and aren't. The monthly API credits that come with Max ($100 or $200) and Team (up to $500, pooled) can't pay for Claude Code: Anthropic's API credits page lists Claude Code under "Doesn't cover." And the API pricing FAQ says new API users "receive a small amount of free credits to test the API," but it doesn't say those credits work in Claude Code, so don't plan around them.

Learning costs nothing, even though practice does. Anthropic's Claude Academy at academy.claude.com (a separate site from our AItomation Academy) runs free self-paced courses, including Claude Code 101, which the Claude Code overview recommends as a next step. You still need a paid plan or API credits to try what a course teaches. Our guide to what a good Claude Code course covers lays out the order to learn it in.

How much does Claude Code cost? Every plan compared

Claude Code costs $20 a month on Pro ($17 a month billed annually), $100 a month on Max 5x and $200 a month on Max 20x, and a Console account charges only for the tokens you use, according to Anthropic's Claude Code page. Team seats cost $25 a month for Standard or $125 for Premium on monthly billing ($20 and $100 on annual billing), and Enterprise charges $20 per seat per month plus usage at API rates, per the pricing page.

Every paid plan includes Claude Code at no extra charge, and it shares your plan's allowance with regular Claude: the pricing page says your work in the terminal and your chats "draw from one pool." The table compares the plans for Claude Code work.

Sources: claude.com/pricing, the Max plan and Team plan help articles, Anthropic's Fable plan article and the API pricing page, checked October 2026.

Pro is the entry point. Anthropic's Claude Code product page calls it "perfect for short coding sprints in small codebases," and the Pro plan article describes a session limit that resets every five hours plus a weekly limit across all models. Fable, the model Anthropic's model docs point at "your hardest and longest-running tasks," sits outside Pro's included usage: on Pro, the pricing page lists Fable under usage credits.

Max 5x and Max 20xinclude five or 20 times Pro's per-session allowance, according to Anthropic's Max plan article, and both bill monthly only. The Claude Code page pitches Max 5x for "everyday use in larger codebases" and Max 20x for "power users with the most access to Claude models." On Max you can use "up to 50% of your weekly usage limits on Fable models at no extra cost," according to Anthropic's Fable plan article, but that share comes out of your regular weekly limits rather than adding to them.

Team needs at least two members, supports up to 150 seats and, per the pricing page, lets you mix seat types on one plan. A Standard seat includes 1.25 times Pro's per-session allowance and a Premium seat 6.25 times, per Anthropic's Team plan article, and Anthropic says Claude Code is included with every Team plan seat. Our Claude Team plan guide covers seats and admin controls in more depth.

Enterprise starts at 20 seats: $20 per seat per month billed annually, plus usage at API rates, according to the pricing page, so the bill rises with the work your people run. On usage-based Enterprise plans, Anthropic's Team and Enterprise article says there are no per-seat usage limits.

Pay-as-you-gomeans signing Claude Code in to a Claude Console account instead of a subscription. Claude Code then "consumes API tokens at standard API pricing," per the Claude Code page, and Anthropic's cost guide explains that Console usage is billed per token, while Team and Enterprise seats draw on a per-seat allowance that resets on a rolling five-hour window and a weekly window. You can cap Console spending with a workspace spend limit. The current Opus costs $4 per million input tokens and $20 per million output tokens, Sonnet costs $2 and $10, Haiku costs $0.10 and $0.50 for prompts up to 100,000 tokens, and Fable costs $10 and $50, according to Anthropic's API pricing page. Older model versions that are still available can cost more.

All of these are US dollar prices, and the pricing page notes that they don't include applicable tax. The Pro plan article says Pro costs "$20 per month (US), with pricing in your local currency where supported," and that monthly pricing varies by region. If you are choosing between the two individual plans, our Claude Pro vs Max comparison covers the chat side of that decision.

What do the 5-hour and weekly limits mean in real hours?

Every paid plan runs two usage meters at once: a session limit on a rolling five-hour window and a weekly limit that resets at a fixed time each week. Anthropic doesn't publish how many hours of Claude Code either meter allows on any plan, so the only reliable number is the one you measure on your own account.

The session limit caps how much you get done in one sitting. When you run out, Claude Code stops and shows the exact reset time, in a message such as You've hit your session limit · resets 3:45pm, and it can warn you earlier, for example once you've used 85 percent, according to Anthropic's Claude Code error reference. The weekly limit caps how many heavy sittings you get. Its reset day and time stay the same "regardless of when you start using Claude," according to the Max plan article, and Settings > Usage on claude.ai shows your next reset.

Both meters fill at the same time. The error reference warns that one burst of heavy activity "can exhaust the weekly allowance before the session window resets," and switching models won't restore access, because both limits apply across all models. When Anthropic launched its newest Opus model on September 22, 2026, it raised five-hour limits on Pro, Max, Team and seat-based Enterprise plans and gave subscribers a limit reset they can save for when they need it.

You will still see hour counts online, such as "40 to 80 hours" a week on Pro. Those figures date from Anthropic's July 2025 announcement of weekly limits, as reported by AIN at the time, and they described two models that Anthropic retired on June 15, 2026, according to its model deprecations page. Anthropic's current help articles on Pro, Max and Claude Code give no hour or prompt counts, so treat any such number as out of date.

Five habits drain both meters faster in Claude Code:

• Leaving one session open all day. Claude Code sends your full conversation with every request, so a one-line question late in a long session still draws usage for everything before it, per the cost guide. Anthropic's Claude Code usage article suggests /clear when you start a new task and /compact when you continue a long one.
• Running Opus on simple jobs. Claude Code defaults to Opus on Pro, Max, Team and Enterprise, according to the model configuration docs. The same usage article calls Sonnet "the right choice for the large majority of coding work" and says Opus "uses meaningfully more of your quota." Type /model to switch, and see our Sonnet vs Opus guide for which jobs need which model.
• High effort on routine work. The cost guide recommends lowering the effort level with /effortfor tasks that don't need deep reasoning.
• Parallel and background work.Running cloud sessions in parallel "consumes more rate limits proportionately," per the cloud sessions docs, and scheduled tasks draw down your plan the same way a live session does.
• A long CLAUDE.md file. Claude Code loads CLAUDE.md at the start of every session, so long instructions cost tokens even on unrelated work. The cost guide suggests keeping it under 200 lines and moving specialized instructions into skills, which load only when needed. Our Claude Code skills guide shows how to set them up.

To see what used your allowance, type /usage in Claude Code. On a paid plan it shows your plan usage bars plus a breakdown of recent usage by skills, subagents, plugins and MCP servers, with suggested fixes, per the cost guide. In the desktop app, the usage ring next to the model picker shows your plan usage for the period. For how the same limits work in regular chats, see our Claude usage limits explainer.

A Claude Code pricing calculator for your own work

The quickest way to price Claude Code for your own work is to measure how much of your weekly allowance one hour uses. Note the weekly percentage before and after a session, divide the change by the hours you worked, then divide 100 by the result: that is roughly how many hours a week your plan covers for that kind of work.

You can read the percentage in three places. Settings > Usage on claude.ai shows progress bars for the session and weekly limits, /usageshows the same plan bars in the terminal, and the desktop app's usage ring shows plan usage next to the model picker.

Anthropic states the plan multipliers per session (the Max 5x plan gets five times Pro's per-session allowance, a Team Premium seat 6.25 times), so treat the weekly math as an estimate. In the example, Max 5x would cover the gap with room to spare. A small or occasional gap costs less to fill with usage credits, which Anthropic's usage credits article says are prepaid, billed at standard API rates, apply to Claude Code, and can carry a monthly spend limit.

To compare a subscription with pay-as-you-go, start from the figures Anthropic publishes: across enterprise deployments, Claude Code costs "around $13 per developer per active day and $150-250 per developer per month, with costs remaining below $30 per active day for 90% of users," according to its cost management guide. At that average, two active days of pay-as-you-go ($26) cost more than a month of Pro ($20), eight days ($104) cost more than Max 5x ($100), and 16 days ($208) cost more than Max 20x ($200).

Those are developer numbers, and lighter jobs such as renaming files or drafting a report script may cost less per day, so measure your own. On a Console account, /usageestimates your session's cost at list price, and the Console's Usage page holds the billing figure, per the cost guide. A subscription wins once you use Claude Code on more than a couple of days a month, as long as its limits cover your work. Pay-as-you-go suits rare use, or a sprint where you would rather pay per token than wait for a reset: Anthropic's Pro and Max help article suggests a Console account for "intensive coding sprints."

Paste a week of notes into Claude with the prompt below, and it will run the calculator for you and recommend a plan:

I want the cheapest Claude plan that covers my Claude Code work.
Use only the numbers below and show your math.

MY USAGE LOG (one line per Claude Code session):
- [date] | [task] | [hours worked] | weekly usage before: [x%] |
  after: [y%] | hit the session limit? [yes/no]
- ...

My current plan: [Pro / Max 5x / Max 20x / Team Standard / Team Premium]
Hours of Claude Code work I expect per week: [number]
Usage credits I paid for last month: [$ amount, or 0]

PLAN FACTS (October 2026, US prices):
- Pro: $20/month, or $17/month billed annually. Baseline allowance.
- Max 5x: $100/month, 5x Pro's per-session allowance.
- Max 20x: $200/month, 20x Pro's per-session allowance.
- Team Standard: $25/seat monthly or $20 annual, 1.25x Pro. 2-seat minimum.
- Team Premium: $125/seat monthly or $100 annual, 6.25x Pro.
- Usage credits: billed at API rates on top of a plan, optional monthly cap.

Calculate:
1. Average weekly usage (%) per hour of work across my sessions
2. Hours per week my current plan covers (100 / step 1)
3. How many times my current allowance I need (expected hours / step 2)
4. One recommendation: stay, add usage credits with a cap, or change
   plan. The multipliers are per session, so call the weekly math an
   estimate.
5. Sessions in my log that look wasteful (very long sessions, simple
   tasks that may not need Opus) and what to change

Get more done per session with prompts and skills that already work

• 500+ copy-paste prompts & workflows
• 35 installable Cowork skills
• Role packs for 6 professions
• New prompts & skills every week, forever

See the Vault

Which Claude Code plan should a non-developer pick?

Start on Pro, measure a week of real work with the calculator above, and move up only when the numbers tell you to. Anthropic's Claude Code page describes Pro as built for short sprints in small codebases, which fits the reports, file clean-ups and small websites most non-developers start with. The table turns the rest into decisions you can make in a minute.

The $80 line is our rule of thumb, and it comes from arithmetic on the published prices. Max 5x costs $80 a month more than monthly Pro (about $83 more than annual Pro), so once your usage credits pass that amount, the same money buys five times Pro's per-session allowance instead of metered overage. If you paid for a year of Pro up front, the Max plan article says that if the remaining balance on your annual Pro plan is greater than the Max price when you switch, you get an account credit for the remaining balance toward future subscription costs, as long as your billing addresses match.

For a two-person business, price Team before you buy Max plus Pro. One Premium seat and one Standard seat cost $120 a month on annual billing ($100 plus $20) or $150 on monthly billing. Max 5x plus Pro costs $117 a month with annual Pro or $120 with monthly Pro. On annual billing, the Team pair costs $3 more and gets 6.25 times plus 1.25 times Pro's per-session allowance, against five times plus one time for Max and Pro, according to the Team plan article. The catch: that Team price is the annual-billing rate, so both seats are billed a year at a time. Fable access matches across the two setups: the Premium seat and Max 5x can each spend up to 50% of their weekly limits on Fable, while the Standard seat and Pro pay for Fable with usage credits, according to Anthropic's Fable plan article.

Before you pay for Claude Code at all, check that your work needs it. If your tasks are documents, research and files rather than scripts or websites, Cowork, which Anthropic is merging into the regular Claude app, may cover them on the same plan. Our Claude Cowork vs Claude Code comparison draws that line, and our Claude Cowork pricing breakdown covers what Cowork costs. On a Team or Enterprise seat, the cost guide notes that one allowance covers Claude chat, Cowork and Claude Code, so you don't pay for them twice.

Want to learn Claude Code without burning your weekly limit on trial and error?

• 1,300+ non-technical professionals learning together
• Claude, Claude Cowork & Claude Code training
• The full 550+ prompt & skill Vault included with Premium
• A weekly live call to get unstuck

Join the Academy

Five ways Claude Code can cost more than your plan price

• An API key on your computer. If the ANTHROPIC_API_KEY environment variable is set, Claude Code uses that key instead of your subscription and bills at API rates, Anthropic warns in its Pro and Max help article. Claude Code asks whether to use the key first, per the setup docs, so approve it only if you mean to pay per token. To see which login is active, type /status and check the API key row, as the error reference suggests.
• Fast mode. Typing /fastmakes Opus up to 2.5 times faster at $8 per million input tokens and $40 per million output tokens, twice the standard Opus price. On a subscription, fast mode bills only to usage credits, "even if you have remaining usage on your plan," according to the fast mode docs.
• Fable on Pro.On Pro, Fable runs on pay-as-you-go usage credits from the start, according to Anthropic's Fable plan article, and those credits bill at standard API rates. When Fable bills that way, the /model picker labels its row "Requires usage credits," and in an interactive session Claude Code asks before a Fable request bills your credits.
• Usage credits with no cap. The usage credits settings offer "Set to unlimited" and automatic top-ups. Set a monthly spend limit the day you turn credits on.
• Work that runs while you're away. With usage credits on, scheduled tasks can keep running past your limit on metered overage, per the scheduled tasks docs. Without credits, new runs are rejected until your usage window resets.

Frequently Asked Questions

Does Claude Code on the web or in the desktop app cost extra?

No. The desktop app includes Claude Code in its Code tab, so you don't install anything else, though it still needs a Pro, Max, Team or Enterprise subscription, per the desktop quickstart. Cloud sessions at claude.ai/code count toward your plan's usage limits, and Anthropic's cloud sessions docs say "there's no separate charge for the cloud VM."

Do scheduled Claude Code tasks cost extra?

Not on their own. Cloud scheduled tasks are available on Pro, Max, Team and Enterprise plans and draw on your subscription usage the same way a live session does, according to the scheduled tasks docs. Each account can start up to 100 scheduled runs an hour, and that hourly cap has no paid overage.

What happens if I hit the limit in the middle of a task?

Claude Code may finish the step it's on with a small, capped "wrapping up" allowance that counts toward your weekly limit, then stops until the reset time it shows you, per the Pro and Max help article. In an interactive session signed in with your subscription, recent versions wait and continue the task after the reset by default, as long as you keep the session open.

Does Claude Code cost the same outside the US?

Not always. The Pro plan article says monthly pricing varies by region, some regions show tax in the price while others add it at checkout, and subscribers who pay through a mobile app should check their app store for local pricing. The Max plan article gives its $100 and $200 prices for web subscriptions only.

Can my company pay for Claude Code through AWS, Google Cloud or Microsoft?

Yes. Claude Code also runs through Amazon Bedrock, Google Cloud's Agent Platform and Microsoft Foundry, per the setup docs. Those providers bill per token to your company's cloud account, and spend controls live in the cloud provider's billing console rather than at Anthropic, according to the cost guide.