Is Claude Free? What Free, Pro and Max Actually Include

Yes, Claude is free, and the free tier is not a trial. On Anthropic's pricing page, read 23 October 2026, Free is listed at $0 with the line "Free for everyone", and it includes chat on web, desktop and mobile, web search, creating files and running code, memory across conversations, connectors to your apps and tools, Artifacts, and Claude Design, Slides and Docs. The same page shows what Free does not include: Claude Code, Claude Science and @Claude are all marked No in the Free column, and the Opus row turns Yes only from Pro upward, which is where the current Opus, Claude Opus 5.5, comes in.

The second half of the query is harder. Spanish and Russian searches about Claude Pro token limits are rising alongside "is claude free", and they are asking a question Anthropic has not answered with a number. The pricing page publishes a relative multiplier and a usage dashboard; it does not publish how many tokens a plan includes. Anyone who gives you a token count for Claude Pro is estimating, measuring their own account, or making it up. This article covers what is published, and says plainly where the page stops.

What the free tier actually gives you

Free is a real product rather than a countdown. Web search is on it, so is code execution for creating and editing files, memory that carries across conversations, connectors, Artifacts, and the three document tools Anthropic lists as Design, Slides and Docs. The feature comparison on the same page marks chat on web, iOS, Android and desktop as available on all four individual variants, and marks voice mode, incognito chats, desktop extensions and skills as available on Free too.

Two limits define the tier. There is no Claude Code, which matters because Claude Code is the terminal product most developers arrive for, and there is no @Claude, Claude Science, Claude Dashboards, or the Chrome and Microsoft 365 integrations. Projects exist on Free but capped at five, while the paid columns list Projects without a cap. If your work is chat, research and the occasional document, that set is enough. If your work is an agent in a terminal, it is not, and no amount of prompt discipline changes which column a feature sits in.

What $17 or $20 buys

Pro is the first paid tier, and it has two prices for the same plan: $17 a month with annual billing, which is $200 up front, or $20 if billed monthly. The list of additions is where most of the value sits for a developer. Pro includes Claude Code and Claude Science, Projects, more Claude models, Claude Dashboards in beta, handing off and scheduling tasks, Claude in Chrome, and Claude for Microsoft 365. The pricing page's FAQ adds a detail the plan cards leave out: Claude Code shares the same usage limits as the rest of your plan, so terminal work and chat draw from one pool.

The annual discount is real, but it is a commitment. Anthropic bills the year once, and the FAQ says switching from monthly to annual credits any unused time toward the first annual invoice. Upgrades take effect immediately with the remaining value credited; downgrades take effect at the end of the current billing period, and chats, projects and files stay with the account either way. Cancelling keeps the plan active until the period ends, and a Pro or Max account moves to Free afterwards. Refunds are generally not available except where the Consumer Terms or local law say otherwise, and the European Economic Area and the United Kingdom get a fourteen-day withdrawal window.

What Max adds, and who it is for

Max starts at $100 a month and is sold as two tiers: 5x or 20x more usage than Pro. Beyond the multiplier it adds higher output limits for all tasks, early access to advanced features, priority access at high traffic times, and $100 to $200 a month in API credits to build on the Claude Platform. The card's framing is "for people who work with Claude all day", and the FAQ is more specific: Max is for people who want longer, uninterrupted sessions.

Whether that is worth it depends on one thing you can actually measure, which is how often you hit a limit. Max is not a different product; it is the same product with a bigger allowance, higher ceilings and a queue position. If you rarely see a limit message, the multiplier buys you nothing you can feel, and the API credits are a separate benefit you should value separately rather than folding into the subscription price.

The three cards are the easy part to photograph and the least stable part to quote. Anthropic's own footnote on the page says usage limits apply, prices shown do not include applicable tax, and price and plans are subject to change at Anthropic's discretion. Read the live page before you budget, not this article and not an image of it.

The token question, answered honestly

Search in Spanish for cuántos tokens incluye Claude Pro, or in Russian for the same question, and you will find pages with confident numbers. Anthropic does not publish one. What the pricing page publishes is a structure: every plan has usage limits that reset on a rolling five-hour session window, and paid plans add weekly limits on top. Your activity across web, desktop, mobile and Claude Code all draws from the same pool. How much you can do "depends on the length and complexity of your conversations, the model you choose, and the features you use, so there's no fixed message count."

What it does publish is a ratio. In the page's FAQ, Pro gives you "at least 5x more usage per 5-hour session than Free", and Max gives you "5x or 20x more usage per 5-hour session than Pro". On Team plans, a standard seat gives more than Pro and a premium seat gives 5x more than a standard seat. Those are relative multipliers against a baseline that is itself not expressed in tokens. A multiplier against an unpublished quantity cannot produce a token count, which is why every "Claude Pro includes N tokens" figure you read is somebody's measurement of their own account, not a specification.

Three practical consequences follow. First, the only ground truth for your account is Settings, then Usage, which the FAQ names as the place to see where you stand. Second, the same nominal plan behaves differently for two people, because a long conversation with a large model and several features enabled consumes more than a short one. Third, when you do hit a limit, the FAQ lists three options: wait for it to reset, move to a higher plan, or, on paid plans, turn on usage credits to keep working at standard API rates. That last option is the only place where dollars re-enter the picture, and it is a rate for extra capacity rather than a conversion of your subscription into tokens. Multiplying a plan's price by an API rate and calling the result your allowance describes a product Anthropic does not sell.

What is published and what is not

It is worth separating the two lists, because most confusion about Claude pricing comes from mixing them. Published: the plan names and prices, the feature matrix, the billing cycles, the relative multipliers, the five-hour window, the existence of weekly limits on paid plans, and the dashboard where you check your own usage. Not published: a token allowance for any plan, a per-plan dollar value for usage, and what a reset actually restores. The pricing page stops where the accounting would begin, which is why the question keeps being asked.

The Claude Code line is the one that catches developers. Claude Code is included in all paid plans rather than sold as a separate subscription, and it draws from the same pool as everything else. The FAQ points heavy coding sessions at pay-as-you-go API credits through a Console account as the alternative, which is a different product with per-token pricing rather than a bigger subscription. That is also the honest answer to "should I buy Max for coding": measure your own usage first, because the multiplier applies to a pool you cannot see the size of in advance.

Beyond the three individual plans

Team and Enterprise are separate tracks rather than bigger versions of Pro. Team is sold per seat for teams of two to 150, with a standard seat at $20 a month billed annually or $25 monthly, and a premium seat carrying 5x the usage of a standard seat at $100 annual or $125 monthly; it adds central billing, single sign-on, domain verification and admin controls. Enterprise is quoted separately, with a self-serve path and a sales-assisted one, and its pricing is a seat price plus usage at API rates rather than a flat subscription. Neither track changes the answer to "is Claude free", since the free tier stays free, but both change what a team should compare before buying seats.

Then there is the question of what to do if you want per-token pricing without a subscription at all. That is the API, and it is a different product from any of the plans above. OrcaRouter hosts models behind one OpenAI-compatible endpoint: 200-plus models, one key, provider list price passed through at 0 percent markup, with automatic failover across providers. Claude Opus 5.5 is on our catalogue at $4.00 per million input tokens and $20.00 per million output tokens, with a one-million-token context and cache reads at $0.20 per million. We host models. We do not sell Anthropic's Claude apps, Claude Code, Claude Tag, or any of the subscription plans described above, and an API key is not a substitute for a Pro or Max subscription.

Choosing in one pass

If you want to know whether Claude is free enough for you, answer three questions in order. Do you need Claude Code or more than five Projects? If yes, Free is out and the decision is between Pro and Max. Do you hit a limit more than occasionally? If yes, compare the 5x and 20x Max tiers against what a week of your own usage shows in Settings, not against a token number from a search result. And do you need per-token billing, central administration or an API? If yes, you are looking at the API, Team or Enterprise, and the individual plans are the wrong comparison entirely.

The short honest version: Claude has a real free tier with real limits, Pro at $17 annual or $20 monthly is where the developer tools switch on, Max from $100 is the same product with a larger allowance, and the token count everyone searches for is not published because the allowance is defined as a multiplier and measured in a dashboard. That is a less satisfying answer than a number. It is the one the pricing page supports.