Claude Pricing: Every Plan, Seat and Token Rate

Claude costs $0 on Free, $20 a month on Pro ($17 on an annual plan), $100 on Max and $200 on Ultra. Teams pay $20 or $100 per seat on annual billing, and the API bills per token with no monthly fee. The real differences between plans are usage limits and Claude Code access, not the chat itself. Prices checked 2026-10-31.

Prices checked 2026-10-31

Every number on this page comes from Anthropic's own pricing page, its help center or its API docs. The list is in the frontmatter. Anthropic says on the page that prices and plans are subject to change at its discretion, and that prices shown exclude tax.

Haiku and Sonnet are on every plan, Free included. Opus starts at Pro. Fable, the top model, is a different story, covered below.

Which Claude page do you need?

Claude pricing is really five separate price lists, and people mix them up constantly. We split them across pages so each one can go deep.

What each individual plan gives you

Free

Free is real, not a trial. You get chat on web, desktop and mobile, web search, file creation and code execution, memory across conversations, connectors, Artifacts, and the Design, Slides and Docs tools. Projects are capped at five. Usage is "limited" every five hours and every week, and Anthropic does not publish a number.

What Free leaves out matters more than what it includes. No Claude Code, no Research mode, no Claude in Chrome, no Microsoft 365 integration, no Opus, and no scheduled or handed-off tasks. Sonnet and Haiku are both available.

Pro

Pro is the plan most people mean when they ask what Claude costs. It adds Claude Code, Research, Projects without the cap, Opus, Claude in Chrome, Claude for Microsoft 365, scheduled tasks and Claude Dashboards (beta). Anthropic's FAQ says Pro gives "at least 5x more usage per 5-hour session" than Free.

The annual option is the only discount on any individual plan. $200 up front works out to about $16.67 a month, which Anthropic rounds to $17. A monthly Pro costs $240 over a year, so annual saves $40. Our full breakdown is on the Claude Pro pricing page.

Max

Max is $100 a month and gives 5x the Pro usage in both the five-hour and weekly windows. It adds priority access at busy times, early access to new features, and $100 a month in API credits for the Claude Platform. Fable, the top model, is included for up to half your weekly limit. Max is monthly only. See Claude Max plan.

Ultra

Ultra is $200 a month with 10x the Pro usage each week and no five-hour limit for verified accounts. It carries $200 a month in API credits. It replaced the old Max 20x plan, and existing Max 20x subscribers moved over automatically at no price change.

How usage limits really work

There are no message counts. Anthropic says so plainly: how fast you reach a limit depends on the model, the length of the chat and the size of your files, so "there's no fixed message count."

Paid plans have two layers.

• A five-hour session limit. The clock starts when you send your first message and the allowance refills in full five hours later.
• A weekly limit. It resets at a fixed time each week assigned to your account, no matter when you subscribed or started using Claude.

Max and Ultra add a third: a separate weekly limit for Fable. Verified Ultra accounts skip the five-hour layer. An Ultra account that is not verified yet gets a temporary limit until Anthropic verifies it automatically.

Everything draws from one pool. Chat on claude.ai, the desktop app, mobile and Claude Code all count against the same limits. This is the single fact that surprises developers most. A long Claude Code session on Pro eats the allowance you would use for chat, and the reverse.

Anthropic also reserves the right to change limits. The help center says it may limit usage with weekly and monthly caps or by model and feature "at our discretion," and that it does not guarantee uninterrupted service. That is normal for a consumer plan, but it means a plan's 5x or 10x multiplier is a ratio to Pro, not a promise of an absolute amount.

What happens when you hit a limit

You have four options, per Anthropic's help center.

1. Wait. Session limits refill after five hours, weekly limits at your reset time.
2. Upgrade. The new plan's limits apply right away.
3. Turn on usage credits. Requests switch to pay-as-you-go at standard API rates and keep going. You prepay a balance, can set a monthly spend cap, and can enable auto-reload.
4. Use a Console account. For intense Claude Code sprints, switch to API billing.

Usage credits are billed separately from the subscription and show up as additional charges. Credits on a plan are separate from the monthly API credits that come with Max, Ultra and Team: those cannot be spent in Claude, Claude Code or Cowork.

One Claude Code detail is worth knowing. If you hit the five-hour limit mid-response, Claude Code may keep working briefly to finish the step in progress. Anthropic calls this the wrap-up allowance. It is small and capped, it counts toward your weekly limit, and on Pro it applies at most once per weekly period. Anthropic says it is rolling out gradually.

Team and Enterprise

Team is for 2 to 150 people. A Standard seat is $20 a month on annual billing or $25 monthly. A Premium seat is $100 annual or $125 monthly and carries 5x the usage of a Standard seat. You can mix seat types, so a team of ten might buy two Premium seats for heavy Claude Code users and eight Standard seats for everyone else.

Team adds what individual plans lack: central billing and admin, single sign-on, enterprise search across Slack and Microsoft 365, project sharing, spend controls at organization and user level, and no training on your content by default. Team plans also include pooled monthly API credits of $20 per Standard seat and $100 per Premium seat, capped at $500 a month for the pool.

Enterprise is $20 per seat per month on annual billing, 20-seat minimum, with usage billed at API rates on top. It adds SCIM, audit logs, a Compliance API, custom data retention and a HIPAA-ready option. Cost scales with the models and tasks your team runs, so the $20 is a floor, not a forecast. Anthropic also sells a separate Education plan for universities; there is no public price.

The API is a different bill

The API has no monthly fee and no relation to your subscription. A Pro plan does not include API usage, and a Console account does not include chat. You pay per million tokens: Sonnet 5.5 is $2 in and $10 out, Opus 5.5 is $4 and $20, Haiku 5.5 is $0.10 and $0.50 for prompts up to 100,000 tokens, and Fable 5.1 is $10 and $50. Batch processing takes 50 percent off. The full table with caching rates is on Claude API pricing, and a cross-vendor view is on LLM API prices.

API accounts also have spend caps by tier: Start $500, Build $1,000 and Scale $200,000 a month. When you reach the cap, requests fail with a 429 until the first of the next month, unless you ask for a higher limit.

Fable: the model with its own rules

Fable is Anthropic's top model, priced at $10 in and $50 out per million tokens on the API. On subscriptions it follows a different rule from everything else.

Anthropic notes that the 50 percent is a cap on Fable's share of your weekly limit, not a bonus on top of it. Fable draws from the same weekly pool and uses it faster than other models. If you never pick Fable, you never pay extra for it. In Claude Code neither Fable model is the default on any plan; you choose one explicitly.

What changed recently

• Max 20x became Ultra. Same $200, now with 10x weekly usage and no five-hour limit for verified accounts.
• Sonnet 5 pricing stayed at $2 and $10. The API docs say the introductory price became standard, and the scheduled rise to $3 and $15 on 1 September 2026 will not happen.
• Fable 5's plan promotion ended on 19 July 2026. Pro and Team Standard seats now use usage credits for Fable. Fable 5.1 was never part of that promotion.
• Opus 5.5 and Sonnet 5.5 cache reads got cheaper. They cost 0.05x base input instead of 0.1x.

Who should pick which

Our take

We did not run usage tests, so we will not tell you how many messages a plan holds. Anthropic does not publish that either. What the documents support is this.

Annual Pro is the cheapest way into Claude Code. If you code with Claude most days and hit limits, the Max jump from $20 to $100 is a 5x step in usage for 5x the price, so the unit price of usage is flat. You are paying for fewer interruptions, not a bulk discount. Usage credits at API rates are the pressure valve, and the monthly spend cap keeps them from becoming a surprise. For a heavy week they may beat a plan upgrade; for a heavy year they will not.

Teams should check the seat arithmetic before defaulting to Premium. A Premium seat at $100 annual equals the price of an individual Max plan, with admin controls added. It also costs 5x a Standard seat for 5x the usage, which makes the question simply whether your heavy users are really that heavy.

And everyone should keep the API in mind as a separate meter. Claude Code on a plan versus Claude Code on API keys can differ a lot. Anthropic's cost doc puts enterprise averages near $13 per developer per active day, or $150-250 a month, which sits above Pro and Max prices for many people. That is an average across enterprise deployments on API billing, not a forecast for you. See Claude Code pricing for the working.

Cheaper or different options

If Claude is more than you need, compare it with the other assistants on cost. Cursor pricing, GitHub Copilot pricing and Codex pricing each have lower entry points, and the AI coding agent pricing comparison lines them up with Claude Code. For the models themselves, see the Claude models page and the Opus vs Sonnet comparison.

Checklist before you pay

• Decide whether you need Claude Code. If yes, Free is out.
• Check whether you need Opus or Fable. Opus needs Pro; Fable on Pro runs on credits.
• Pick annual only if you are sure; annual refunds are not standard, because Anthropic's payments are "generally non-refundable."
• Turn on usage credits with a low monthly cap, or leave them off if you want a hard stop.
• If you also build with the API, claim the Max or Ultra monthly API credit and link one Console organization. You can link only one, and you cannot change it yourself.
• Remember that ANTHROPIC_API_KEY in your shell makes Claude Code bill the API instead of your plan.