The PromptsRush journal
AI ToolsClaude Pricing Explained: Pro, Max, Ultra, Team, Enterprise
Every Claude plan, updated for November 2026: Free, Pro, Max, the new Ultra tier, Team seats and Enterprise. What each tier buys, which include Fable and API credits, and which fits how you work.

Advertisement
Most people overpay for Claude in one of two directions: they sit on the Free plan hitting limits every afternoon, or they buy the $200 tier when the $20 one would have covered them. The right plan follows from one question — how many hours a day is Claude actually doing work for you? Under an hour, Pro is enough. Most of your workday, you want Max. If Claude Code runs long autonomous sessions while you sleep, that is what the new Ultra tier is for.
This guide covers every tier as it stands in November 2026 — Free, Pro, Max, Ultra, Team's two seat types and Enterprise — with every price checked against Anthropic's pricing page on November 3, 2026. Since we first published it in July, the lineup has changed in ways that matter for your bill: Max 20x became Ultra and dropped its 5-hour cap for verified accounts, access to Claude Fable — Anthropic's top model line — now depends on your tier, and three plans include monthly API credits. Anthropic adjusts plans and limits often, so confirm the numbers on the official page before you subscribe.
What Changed Since Mid-2026
On October 30, 2026, Anthropic renamed its two Max tiers and changed how the top one is metered. Prices did not move, and existing subscribers were switched over without doing anything. Together with the other items the pricing page now marks as new, the changes come down to four things:
- Max 5x is now just Max. It is still $100 a month with 5x Pro usage per 5-hour window and per week, and it now includes Fable up to half your weekly limit plus $100 a month in Claude API credits.
- Max 20x is now Ultra. It is still $200 a month, but verified accounts have no 5-hour limit, the weekly allowance is listed as 10x Pro, and it includes $200 a month in API credits.
- Fable access is tiered. Free has none, Pro pays for it with usage credits at API rates, and Max, Ultra and Team premium seats include it up to half the weekly limit.
- Max, Ultra and Team include monthly API credits. Team pools them across seats, up to $500 a month for the whole workspace.
Every Claude Plan at a Glance
| Plan | Price | Usage | Fable | Best for |
|---|---|---|---|---|
| Free | $0 | Limited, resets every 5 hours and weekly | No | Trying Claude, light chat use |
| Pro | $20/month, or $200/year ($17/month) | At least 5x Free per 5-hour window | Paid with usage credits | Daily individual use, focused Claude Code sessions |
| Max | $100/month, monthly only | 5x Pro per 5 hours and per week | Up to half the weekly limit | Heavy daily work, Claude Code open all day |
| Ultra | $200/month, monthly only | 10x Pro per week, no 5-hour limit once verified | Up to half the weekly limit | Long autonomous runs, parallel agents |
| Team (standard seat) | $25/seat monthly, $20 annual | More than Pro | Not included | Most people on a team of 2–150 |
| Team (premium seat) | $125/seat monthly, $100 annual | 5x a standard seat | Up to half the weekly limit | The engineers who live in Claude Code |
| Enterprise | $20/seat/month, billed annually (20-seat minimum), plus usage at API rates | Metered at API rates | Billed as usage | Compliance, SCIM and audit requirements |
Plans differ in two ways: how much you can use, and which models you get. Opus, Sonnet and Haiku run on every paid plan, but Fable is a tier feature. So the decision comes down to two questions — how many hours a day Claude works for you, and whether that work needs the top model.
Which Claude Plans Include Fable
Fable is where the plans diverge most. The pricing page breaks it down like this:
| Plan | Fable access | Other models |
|---|---|---|
| Free | Not available | Sonnet, Haiku |
| Pro | Pay as you go with usage credits, at standard API rates | Opus, Sonnet, Haiku |
| Max and Ultra | Included, with its own weekly cap of up to half your weekly limit | Opus, Sonnet, Haiku |
| Team | Included on premium seats, up to half the weekly limit | Opus, Sonnet, Haiku |
| Enterprise | Available, billed as usage at API rates | Opus, Sonnet, Haiku |
On Pro, Fable runs on usage credits, which bill at standard API rates — Fable 5.1 lists at $10 per million input tokens and $50 per million output tokens — so one long Fable session can cost more than the subscription itself. If you reach for Fable every day, Max is the cheaper way to get it.
Most coding and writing doesn't need Fable at all. Opus is included on every paid plan, and our breakdown of when to pick Opus 5.5 over Fable 5.1 covers the cases where the top model actually earns its price.
The Free Plan: Useful, but Not for Coding
Free covers more than chat. It includes web search, file creation with code execution, memory, connectors, Artifacts, Skills, voice mode, Claude Design, Slides and Docs, and up to five projects. Usage resets on a rolling 5-hour window with a weekly cap on top, and the models are Sonnet and Haiku — no Opus, no Fable.
What Free leaves out is Claude Code, Research and unlimited projects. Claude Code starts at Pro; if you want to run it without a subscription, we wrote up every legitimate route — local proxies, free provider tiers and local models — in how to use Claude Code for free.
Free is for evaluating, not working. If you've hit the limit twice in one week, $20 has already paid for itself.
Pro at $20: The Default Answer for Individuals
Pro is $20 per month, or $200 billed annually — $17 a month. It gives you at least 5x the Free plan's usage per 5-hour window and the full individual toolkit: Claude Code, Opus, Research, Claude Science, unlimited projects, handed-off and scheduled tasks, Claude Dashboards (beta), Claude in Chrome, and the Microsoft 365 and Outlook integrations. Fable is available on usage credits; monthly API credits are not included.
Pro fits anyone using Claude daily for writing, analysis and coding sessions measured in minutes per task rather than hours per day. It comfortably covers a developer who reaches for Claude Code several times a day for focused work — a refactor here, a test suite there — the pattern behind our Opus 5.5 coding prompts.
It breaks down on long agentic sessions and daily Fable use. Claude Code working through a big codebase burns the 5-hour allowance quickly, and every Fable message is billed on top of the subscription. When either becomes routine, that is the signal to move up, not a reason to ration yourself.
Max vs Ultra: The Usage Ladder
Max and Ultra are the two individual tiers above Pro, and both bill monthly only. Each adds Fable up to half your weekly limit, early access to new features, and priority access at high-traffic times. The difference is how the allowance is metered:
- Max ($100/month) — 5x Pro usage per 5-hour window and per week, plus $100 a month in API credits. This is the plan for people whose workday runs through Claude: several long Claude Code sessions a day, big-context research, heavy document work. Most full-time developers who feel squeezed on Pro land here and stop feeling squeezed.
- Ultra ($200/month) — no 5-hour limit for verified accounts, 10x Pro usage per week, and $200 a month in API credits. It is built for work that runs in long, uninterrupted stretches: overnight autonomous runs, multi-agent workflows in Claude Code, or Claude as the engine of a one-person business.
Beyond double the weekly allowance and credits, Ultra's real difference is the missing 5-hour window. On Max, a long agentic session can stall mid-task when the window runs out; on a verified Ultra account, the weekly limit is the only ceiling. Anthropic verifies Ultra accounts automatically from account and payment history, and new subscribers keep a temporary 5-hour limit until that happens. If Settings > Usage still shows a "Current session" bar, your account isn't verified yet.
Read the weekly number rather than the old name: the tier once sold as Max 20x now lists 10x Pro per week, double Max's 5x. If you don't already know you need Ultra, you don't.
Pro tip: Upgrade reactively, not aspirationally. Start at Pro, work normally for two weeks, and let the limit warnings tell you whether Max is a real need or an imagined one. Max and Ultra bill month to month, so you are never locked into a tier you've outgrown — but the $80 jump from Pro to Max is real money.
Monthly API Credits on Max, Ultra and Team
The most underrated change in the lineup: three plans now come with credits for the Claude API, Anthropic's developer platform. These credits are separate from your chat and Claude Code limits and do not stretch them.
| Plan | Monthly API credits |
|---|---|
| Free, Pro, Enterprise | None |
| Max | $100 |
| Ultra | $200 |
| Team | $20 per standard seat and $100 per premium seat, pooled and capped at $500 for the workspace |
The credits work on the Claude API (Messages and Message Batches), the Console Playground, Claude Managed Agents and the Agent SDK, with any Claude model on the Claude Platform. They do not cover interactive Claude Code in the terminal, IDE, desktop app or web, extra usage inside the Claude apps, or Claude on Amazon Bedrock, Google Cloud Vertex AI or Microsoft Foundry. Headless claude -p runs and Agent SDK scripts draw on the credits only when they use an API key from the Console organization you linked; signed in with your plan, they count against plan limits instead.
The rules are strict. You claim the credits on claude.ai in a web browser by linking one Console organization — under Settings > Billing on Max and Ultra, or Organization settings > Billing for Team owners — and only after seven days on an eligible plan. Unused credits expire at the end of each billing cycle; nothing rolls over. When they run out, API requests stop unless the organization has purchased credits or auto-reload, and nothing is ever charged to your Claude plan.
What $100 buys depends on the model. At list prices it covers about 25 million input tokens plus 5 million output tokens on Sonnet 5.5, half that on Opus 5.5, or a fifth of it on Fable 5.1. For a developer already paying for side-project API usage, that offsets most or all of the Max subscription. If you never touch the API, the credits are worth nothing to you — judge the plan on usage alone.
Pro tip: Link your Console organization as soon as the seven-day wait is over. Credits arrive each billing cycle and expire unused, so a Max subscriber who never links an organization is leaving $100 a month on the table.
Team: Two Seat Types, One Decision
Team covers 2 to 150 people and comes in two seat types you can mix inside one workspace:
| Seat | Monthly | Annual | What it buys |
|---|---|---|---|
| Standard | $25 per seat | $20 per seat | Every Claude feature with more usage than Pro, plus $20 a month toward the pooled API credits |
| Premium | $125 per seat | $100 per seat | 5x a standard seat's usage, Fable up to half the weekly limit, and $100 a month toward the pooled API credits |
Every seat gets the workspace layer: Claude Code and Claude Cowork, Claude Design, Slides and Docs, Claude Motion and Dashboards (both in beta), Claude Science, Claude Tag, Microsoft 365, enterprise search, central billing, SSO, admin controls for connectors, organization-wide skills and usage analytics. Team also excludes your content from model training by default; on individual plans, training is a setting you opt out of.
The mixing is the point. A 20-person company doesn't need 20 premium seats — it needs premium for the four engineers running Claude Code all day and standard for everyone else. That blend costs $900 a month on monthly billing (4 × $125 + 16 × $25) or $720 a month billed annually (4 × $100 + 16 × $20), and it earns the full $500 a month in pooled API credits.
The same people on individual plans — four on Max, sixteen on monthly Pro — also cost $720 a month, with no SSO, no central billing, no admin controls, and $400 in API credits split across four separate accounts. Annual Team matches that price and adds the management layer. And with a two-seat minimum, there is no headcount reason left to share one login, which breaks the terms, memory and projects anyway.
Enterprise: $20 a Seat Plus Usage
Enterprise has a published price: $20 per seat per month, billed annually, with a 20-seat minimum, plus all usage billed at API rates. You can sign up self-serve or go through sales. The seat fee buys a governance layer on top of everything in Team: admin-set spend limits per user and organization, role-based access with fine-grained permissions, SCIM provisioning, audit logs, the Compliance API, custom data-retention controls, network-level access control, IP allowlisting, domain capture, inference hooks, a HIPAA-ready offering and Claude Security (beta). The sales-assisted route adds invoicing on net terms, AWS Marketplace purchasing, consultation, trials and incentives on committed spend.
The economics run opposite to Team. Enterprise is cheap for light users — the 20-seat floor costs $400 a month before usage — and gets expensive for heavy ones, because every token is metered at API rates instead of drawn from a flat allowance, and Enterprise plans don't receive monthly API credits. Size your heaviest users before you sign; our Claude Code API pricing breakdown shows how fast per-developer coding spend adds up. Choose Enterprise when your security team's checklist demands it, not for the price.
Claude API Pricing for Developers
If you build on Claude rather than chat with it, you pay per token on the Claude Platform. Current list prices, per million tokens:
| Model | Input | Output | Cache read | Batch (input / output) |
|---|---|---|---|---|
| Fable 5.1 | $10 | $50 | $0.25 | $5 / $25 |
| Opus 5.5 | $4 | $20 | $0.20 | $2 / $10 |
| Sonnet 5.5 | $2 | $10 | $0.10 | $1 / $5 |
| Haiku 5.5 (prompts up to 100K tokens) | $0.10 | $0.50 | $0.01 | $0.05 / $0.25 |
| Haiku 5.5 (prompts over 100K tokens) | $0.50 | $2.50 | $0.05 | $0.25 / $1.25 |
Opus 5.5 is the rare upgrade that costs less than its predecessor: $4/$20 against Opus 5's $5/$25, with cache reads 60% cheaper. Our Opus 5.5 pricing guide works through the caching break-evens and fast-mode costs in detail. Every current model except Haiku 5.5 bills its full 1M-token context window at the standard rate; Haiku 5.5 switches to its higher price once a prompt passes 100,000 tokens.
- Prompt caching — 5-minute cache writes cost 1.25x the input price and 1-hour writes 2x. Reads cost 2.5% of input on Fable 5.1, 5% on Opus 5.5 and Sonnet 5.5, and 10% on other models; our guide to Claude prompt caching covers where to put the breakpoints.
- Batch API — 50% off input and output for asynchronous jobs, and it stacks with caching.
- Fast mode — up to 2.5x faster output on Opus 5.5 for $8/$40 per million tokens, double the standard rate.
- US-only inference — a 1.1x multiplier on all token prices when you pin inference to the US.
- Tools — web search costs $10 per 1,000 searches and web fetch adds nothing beyond tokens. Code execution is free alongside web search or fetch; otherwise each organization gets 50 free container-hours a day, then $0.05 per hour. Claude Managed Agents add $0.08 per session-hour of active runtime.
Which Plan for Claude Code?
Claude Code is where usage concentrates, and it comes with every paid plan — not Free. It shares one pool with your chats, so a heavy day in the terminal eats into the same 5-hour and weekly limits.
| Your pattern | Right plan |
|---|---|
| Occasional focused tasks, a few sessions a day | Pro |
| Open all workday; multi-file agentic runs | Max |
| Long autonomous runs that hit the 5-hour wall; parallel agents | Ultra |
| Bursty — intense weeks, quiet weeks | Pro, with usage credits for the spikes |
| Headless scripts and automation | An API key from a linked Console organization — Max and Ultra credits cover claude -p and the Agent SDK |
| A team of engineers sharing infrastructure | Team with premium seats |
A subscription and the API aren't mutually exclusive. Plenty of developers keep Pro or Max for interactive work and run production automation on API keys, which keeps personal exploration and metered business usage cleanly separated — and on Max or Ultra, the monthly credits now pay for the first slice of that automation. If what you mostly want is Claude Code on a timer, our Claude Routines explainer covers the interval floor and daily run caps before you upgrade for it.
Annual vs Monthly Billing
Annual billing is available on Pro and Team only. It takes Pro from $240 to $200 a year — $17 a month — and Team seats from $25 to $20 (standard) and from $125 to $100 (premium). That's a 17–20% discount for committing, worth taking once a plan has survived two or three months of real use. Max and Ultra are monthly only, and Enterprise is annual only.
Skip annual on day one. The whole point of the ladder is that your first guess about your own usage is usually wrong.
Mistakes We See People Make
- Buying a bigger plan to fix a prompting problem. If Claude's answers miss, more usage buys you more misses. Fix the workflow first — our Fable 5 prompt templates are a cheaper upgrade than a new tier.
- Assuming every plan runs the top model. Plans now differ in model access, not just volume. Free has no Fable, Pro pays for it per token, Max, Ultra and Team premium seats include it, and Enterprise bills it as usage.
- Never claiming the API credits. Max, Ultra and Team credits require linking a Console organization and expire every billing cycle. Unclaimed, they are worth nothing.
- Expecting API credits to stretch Claude Code. They don't touch interactive Claude Code or your in-app limits. Plan usage and API credits are separate pools.
- Sharing one account across a team. It violates the terms, breaks memory and projects, and costs more in confusion than Team seats cost in dollars.
- Ignoring the API crossover point. At sustained heavy usage a subscription is a bargain; at sporadic usage it's a gym membership. Check your actual pattern quarterly.
The Verdict
Pro at $20 is still the right answer for most individuals and the place everyone should start. Move to Max when limit warnings become a pattern rather than an event, or when you want Fable without paying per token. Ultra is for the small tail whose work runs in long, uninterrupted sessions — the missing 5-hour cap is what you're paying for. Teams of two or more belong on Team with mixed seats, and Enterprise exists for organizations whose real question is governance, not capacity.
If you build on the API, count the credits: $100 a month on Max and $200 on Ultra can cover a side-project API bill, which shifts the math more than any usage multiplier. Whichever tier you pick, the best upgrade per dollar is usually not the next tier up — it's using the one you have better.
Keep Reading
For how Fable stacks up against OpenAI's flagship, see GPT-6 Astra vs Claude Fable 5.1; for Anthropic's new agent app and what it costs, read Claude Duet explained. Or browse all guides and prompts on PromptsRush.
Frequently Asked Questions
10 questions answered


