Claude Plans and Limits: What the Free Tier Really Gets You

#01What does each plan include?

Anthropic's pricing page lists four individual plans. Here's what matters for marketing work (Anthropic).

All four plans include web search, memory, Skills, artifacts, file creation with code execution, incognito chats and connectors. The pricing page notes that prices don't include tax and that plans can change at Anthropic's discretion.

There are also Team and Enterprise plans for organisations. These add admin controls, shared projects and, on Enterprise, features such as enterprise search. If your company already has one, use that account for work rather than a personal plan, because your admin's settings and data terms apply.

#02How do usage limits work?

Anthropic describes usage limits as your "conversation budget". There are two kinds (Anthropic).

The five-hour limit starts when you send your first message and resets in full five hours later. If you hit it early, you wait until the window ends. Verified Ultra accounts don't have this limit.

Weekly limits apply across all models and reset at a fixed time each week assigned to your account. Some plans also have a separate weekly limit for Fable.

You can see your current use and the time left in Settings > Usage. Anthropic also notes that it may adjust limits to manage capacity.

How quickly you use your budget depends on the model you choose, the length of your chats, the size of your files and the tools you use. Opus uses more than Sonnet, which uses more than Haiku. A chat with a 200-page PDF attached costs far more per message than a short exchange. Research, web fetches of long pages and connectors all add up too.

On paid plans, you can turn on usage credits so that work continues when you hit a limit, rather than stopping. That's useful for deadline days, but set a budget so it doesn't surprise you.

#03What does the free plan really get a marketer?

More than many people expect. On the free plan you can:

• draft and edit copy with Sonnet, one of Anthropic's current models;
• search the web with cited answers;
• create documents, spreadsheets and slide decks and run code to analyse a CSV;
• build artifacts such as a simple calculator or landing page mock-up;
• keep up to five Projects with your brand guide and reference material;
• create and use Skills, and connect apps through connectors.

What you'll miss are Opus for harder work, Research for long, multi-source reports, scheduled tasks, Claude in Chrome and the Office add-ins. You'll also hit limits faster, especially with long documents.

Our view: use the free plan for the first four lessons of this track. If you find yourself hitting limits most days, or you need Research or scheduled tasks, Pro pays for itself quickly for anyone whose time is billed or measured. Max and Ultra are for people who use Claude all day, such as content leads running many drafts or analysts working with large files.

#04How do you make your limits last longer?

Anthropic's own usage tips are practical, and most also improve your results (Anthropic).

Use Projects for material you reuse. Content in Projects is cached and counts less against your limits when reused. Upload your brand guide once to a project rather than attaching it to every chat. Anthropic notes that caches expire after a period of inactivity, so the first message after a long break counts the content in full again.

Batch similar requests. Instead of asking for one headline, then another, then another, ask for twenty in one message with clear criteria. It's cheaper and usually better, because Claude can compare options against each other.

Edit your prompt before sending. If a reply misses the mark, edit and resend your original message rather than adding a correction. Every follow-up carries the whole conversation with it.

Start new chats for new tasks. Long conversations get more expensive with every message. When you switch topic, start fresh. Claude's automatic context management can summarise older messages when a chat gets long, but Anthropic notes that this uses more of your usage too (Anthropic).

Turn off connectors you don't need. Anthropic says tools and connectors are token-intensive. If you've connected Drive, Slack and your CRM, but today's task is writing ad copy, switch them off for that chat.

Pick the right model. Use Haiku for simple, high-volume tasks such as categorising a list of keywords. Use Sonnet for most writing. Save Opus for tasks where its extra judgement shows.

Keep project instructions short. Anthropic advises keeping project instructions concise and focused on essential context, and removing project files you no longer use.

#05How do you decide whether to upgrade?

Answer three questions honestly.

1. Do I hit limits more than twice a week? If so, the friction is costing you more than the subscription.
2. Do I need a paid-only feature? Research, scheduled tasks, Chrome and the Office add-ins are the usual reasons. Lessons 7, 12 and 14 cover these.
3. Will my company pay? Many teams now budget for AI tools. If you're on a team of three or more, a Team plan adds shared projects and admin controls, which matter once you're building shared prompt libraries.

If you're unsure, try Pro monthly at $20 for a month while you work through this track. Track what you used that the free plan doesn't include. If the list is short, downgrade.

#06What about the API and Claude Code?

You'll see references to the Claude API, Claude Code and API credits on the pricing page. The API is for developers building Claude into products and is billed by usage, separately from your app subscription. Max and Ultra now include monthly API credits, according to the pricing page. Most marketers won't need either, but it's worth knowing they exist if your team builds internal tools. Lesson 13 on MCP explains where developers and marketers meet.

#07From MentionX data: why prompt type matters more than planOriginal data

Method. Read-only aggregate queries over MentionX scans from 22 August to 29 October 2026 (UTC): 2,217 engine runs (2,154 completed) across 18 anonymised projects and six engines.

Upgrading gives you more usage, not better visibility. In MentionX's aggregates, what the buyer asks shapes whether a brand appears far more than anything else:

Our view: when you use Claude to research your market, spend your limits on the question types that matter, such as comparisons and recommendations. These are where brands are most and least visible.

Limits. A convenience sample of mostly B2B software projects, run through engine APIs, which can behave differently from the consumer apps. Perplexity, Copilot and DeepSeek figures come from a single day (26 October 2026), so treat them as a snapshot, not a trend.

#08Worked example: Ledgerline's plan decision

Ledgerline is a fictional accounts-payable tool. Figures are illustrative.

Ledgerline's demand-gen manager, Tom, starts on the free plan. In week one he hits the five-hour limit twice, both times after attaching the same 80-page analyst report to new chats.

He moves the report into a project, which helps immediately. But by week two he wants Research for a competitor overview and a scheduled weekly scan, neither of which the free plan includes. He switches to Pro monthly and keeps a simple log of what he uses.

After a month, the log shows Research used six times, two scheduled tasks running and Opus used for most analysis. He keeps Pro and recommends the team look at a Team plan once three people are using Claude daily. Figures are illustrative.

#09What should you check before the next lesson?

Open Settings > Usage and note when your weekly limit resets. Create one project and upload your brand guide and two example articles; you'll build on it in lesson 4. If you're tracking AI visibility as well as using Claude for work, our guide to tracking AI traffic in GA4 explains how to see visits that come from assistants such as Claude.

Next: prompting Claude for marketers. Or return to the Claude for Marketers track.

#10Why do limits feel different from day to day?

Because the same number of messages can cost very different amounts. A day of short headline tweaks on Sonnet barely touches your budget. A day of long PDFs, Research runs and Opus analysis can use it up by lunchtime. If you plan heavy work, do it early in a five-hour window, keep it in one project so cached content counts less, and save quick tasks for later. Anthropic also notes that limits apply across every surface you use, so time in Chrome or the Excel add-in comes out of the same budget as chat.

#11How do you build the case for a paid plan at work?

If you want your company to pay, make the case with evidence rather than enthusiasm. Keep a simple usage diary for two weeks on the free plan. For each task, write down what you asked Claude to do, roughly how long it would have taken you without it and whether you hit a limit or needed a paid feature.

At the end of the fortnight, you'll have three useful numbers: how often limits stopped you, how many tasks needed a paid feature and an honest estimate of time saved. Your own estimates are fine, as long as you label them as estimates. Finance teams are used to approving tools on that basis, and it's far more persuasive than a generic claim that AI saves time.

Add a short note on risk. Which data will you put into Claude, and which won't you? Who reviews AI-assisted work before it's published? If your company has an AI policy, show that your use fits it. If it doesn't, offer a one-page draft. Approvals often stall on governance questions rather than cost.

For teams, compare individual Pro subscriptions with a Team plan. Team plans add shared projects, which let colleagues view or edit the same project, and admin controls over features and connectors (Anthropic). Those matter once you have a shared brand-voice project and prompt library, which you'll build in the next two lessons. Check the pricing page for current seat prices, because they change.

Our view: start small. One paid seat for the person who'll build the shared setup is often enough to prove the value, and it gives everyone else a working project to use when the team plan arrives.

#12Which paid features matter most to marketers?

When you decide on a plan, list the workflows you'll actually use, then check which plan includes them. Based on Anthropic's help centre at the time of writing:

If none of the paid-only features appear in your weekly work, the free plan may be enough to start. If two or more do, the time saved usually makes the case for you. Check the current plan pages before buying, because features move between plans.

#13Common questions

Is Claude free? Yes. The free plan includes chat, web search, file creation, artifacts, connectors and up to five Projects, with usage limits.

How much is Claude Pro? $17 a month billed annually or $20 billed monthly, according to Anthropic's pricing page in November 2026, excluding tax.

When do Claude limits reset? The session limit resets five hours after your first message. Weekly limits reset at a fixed time assigned to your account.