The short answer: Claude's plans run from $0 (Free) to $200/mo (Max 20x), and every tier above Free sells the same models with a different usage allowance. Pro at $20/mo - about $17/mo if you pay annually, or $200 for the year - is the plan most writers should buy. Max 5x at $100/mo and Max 20x at $200/mo add capacity and priority, not features: the context window stays at 200K tokens on all of them, and nothing new unlocks. Team sits at $25 per seat per month, or $20 per seat on annual billing, and is a different purchase entirely.
Claude is the strongest prose writer in our 30-day blind benchmark, with the joint-fastest editing time in the set - so the pricing question is not whether the writing is good. It is whether the free tier is enough, and if not, how far up the ladder you actually need to go. Most people stop one rung lower than they think.
Every Claude plan in one table
Prices below were verified in September 2026. Note the billing period on each line - the annual discount is the largest variable in the whole table.
| Plan | Monthly billing | Annual billing | What it changes |
|---|---|---|---|
| Free | $0 | - | Rolling usage budget, roughly 5-hour session limit, no credit card |
| Pro | $20/mo | ~$17/mo ($200/yr) | Full model lineup, much higher limits, Claude Code included, 200K context |
| Max 5x | $100/mo | - | 5x the Pro usage allowance, priority access at busy times |
| Max 20x | $200/mo | - | 20x the Pro usage allowance, highest priority access |
| Team | $25/seat/mo | $20/seat/mo | Shared workspace and admin controls, billed per seat |
The multiplier ranges are the ones Anthropic's own help documentation gives for the Max tiers, relative to Pro. We did not reverse-engineer them into message counts - see the section below for why.
Max is a capacity upgrade, not a feature upgrade
This is the most common misunderstanding about Claude pricing, and it is an expensive one, because the gap between Pro and Max 5x is $80 a month.
Buying Max does not get you a better model, a longer context window, or a feature Pro lacks. The context window stays at 200K tokens on Pro, Max 5x and Max 20x - it does not grow as you pay more. What grows is how much you can send through the model before you hit a limit, plus how far ahead of other users you sit when demand is high.
So the test for Max is not "do I want better output." It is "do I get blocked by Pro's limits often enough that being blocked is costing me more than $80 a month." For most writers, editing time - not model calls - is the constraint on output, and in our benchmark Claude's drafts needed about 8 minutes of editing per 1,500-word piece. That is time Max cannot buy back. A writer paying $100/mo for Max to "write more" is usually solving the wrong bottleneck.
Where Max does earn its price
Two profiles justify it, both volume-driven:
- Long agentic sessions with Claude Code. When a single task runs for hours and consumes usage continuously, capacity is the binding constraint and Max removes it.
- All-day interactive use. If you hit the ceiling most days, the higher multiplier is the point of the plan.
If neither describes your week, Pro is the correct purchase and the $80 difference is better spent elsewhere.
Usage is measured in multipliers, not messages
Claude does not publish a per-plan message count, and the plans are not comparable by counting prompts. What the tiers actually scale is a usage allowance measured in multiples, with a session limit that resets on a rolling basis - roughly every five hours - plus a longer weekly ceiling on top.
The practical translation, as Anthropic describes it: Pro's allowance is more than 5x the free tier's, and the two Max tiers are 5x and 20x of Pro respectively. Two consequences follow, and both matter when you are deciding what to buy:
- A "message" is not a fixed unit. A short question and a 20,000-word document analysis consume very different amounts of allowance, so any table quoting "X messages per 5 hours" is an approximation that will not match your own pattern.
- The limits bind in bursts. Because the session ceiling resets on a rolling window, a heavy morning can lock you out in the afternoon while the same volume spread across the day would not have triggered it.
So the honest answer to "how many messages do I get on Pro?" is: enough for sustained daily writing, not enough for all-day agentic use. We could not verify a precise message count for any tier as of September 2026, because Anthropic does not publish one.
What the free tier actually gives you
The free tier is genuine - no credit card, no expiry, and the same writing quality that makes Claude worth considering in the first place. What it does not give you is headroom: the usage budget resets on a rolling basis of roughly five hours, which in practice means a handful of short drafts a day before you are waiting.
It is the right place to start and a poor place to stay. Two weeks of trial use tells you what a pricing page cannot - whether Claude's prose style fits your voice. If you write daily, the caps stop being an inconvenience and become a scheduling constraint in the first week or two, which is when Pro starts paying for itself. For what else is available free, see free AI writing tools.
Pro at $20/mo: the plan most writers should buy
Pro is the plan we would start on, and at the annual rate it is one of the cheaper paid tiers in the entire market.
- $20/mo on monthly billing - cancel any time, $240 over a year.
- About $17/mo on annual billing - $200 charged for the year, saving roughly $40.
That is a smaller annual discount than most competitors offer in percentage terms, and worth knowing before you commit: the saving is real but modest, so paying monthly for the first couple of months costs you very little and buys you an exit. Take the annual rate once you know you are staying.
What Pro buys is the full model lineup, a substantially higher usage allowance than Free, priority at busy times, Claude Code included at no extra cost, and a 200K-token context window - enough to paste a brand guide, a long report, or several chapters of a draft into one conversation. In our blind benchmark Claude produced the highest-scoring prose we recorded, with about 8 minutes of editing per finished piece - level with Jasper, ahead of ChatGPT's roughly 10 and Copy.ai's roughly 15. The method and its limits are on how we test, with raw results in our 2026 benchmark.
One correction worth making, because it is repeated widely: Pro's context window is 200K tokens, not a million. Larger context windows exist on Anthropic's developer API under specific conditions, but that is a different product with per-token billing, not a feature of the $20 consumer plan.
Team seats: when the $25-per-seat maths works
Team is $25 per seat per month, or $20 per seat on annual billing, and it is the one Claude plan priced per person rather than per account.
For a five-person team on annual billing, that is $100/mo for the whole group - cheaper per head than five separate Pro subscriptions at $17/mo annual, with admin controls and a shared workspace layered on top. For a two-person team, it is $40/mo on annual billing against $34/mo for two Pro seats, so you are paying a small premium for the admin features. Below three seats the maths does not favour Team; above five it does, provided everyone actually uses Claude.
One caveat we cannot resolve from public pricing: what a Team seat's usage allowance is relative to Pro, and how per-seat limits behave under heavy use. We could not verify that as of September 2026, so if your team runs long sessions daily, confirm it with Anthropic before committing to a year.
Buy in a browser, not inside the mobile app
This is the easiest way to overpay for Claude. Subscriptions purchased inside the iOS or Android app carry app-store commission, so the same Claude plan bought through the mobile app is priced above the web price. The plan is identical; only the channel differs.
The fix: subscribe on claude.ai in a browser, then sign in to the mobile app with the same account. You keep the web rate and full mobile access. Check both prices before you confirm - thirty seconds of effort on a subscription you may keep for a year.
Claude against ChatGPT at the same $20
Both entry paid tiers cost $20/mo, which makes this the comparison most readers actually face.
| Option | Price | Editing time per piece | Note |
|---|---|---|---|
| Claude Pro | $20/mo, ~$17/mo annual ($200/yr) | ~8 min | Highest prose score in our blind test; no SEO tooling |
| Claude Free | $0 | ~8 min | Same writing quality, much lower limits |
| ChatGPT Plus | $20/mo | ~10 min | Broader assistant, image generation, wider integrations |
| ChatGPT Go | $5/mo | - | A middle tier between Free and Plus |
| ChatGPT Team | $25 per person/mo | - | Per-seat, like Claude Team |
| Claude Max 5x | $100/mo | ~8 min | Capacity only - same prose, more headroom |
The pattern is consistent with our testing: Claude wins on prose and on how little cleanup it needs, ChatGPT wins on breadth. Note the first two rows - the free tier writes as well as the paid one, which is the strongest argument for trying before you buy. Our ChatGPT vs Claude comparison covers the head-to-head, and is ChatGPT worth paying for handles the other side.
What Claude does not include at any price
This is the constraint that pushes some writers off the ladder entirely, so it belongs in a pricing guide.
Claude has no SEO tooling, no content templates, and no brand-voice management system, and paying more does not add them. Max does not include an SEO suite. Team does not include a brand-voice library. What you get at every tier is an unusually good general writer, and that is the whole product.
If your work is SEO content at volume, the honest advice is to keep Claude for drafting and add a specialist tool for the parts it cannot do - keyword scoring and AI-search visibility tracking. The current reference point there is Writesonic at $79/mo on annual billing, which has repositioned as a GEO (AI search visibility) platform; note that this is annual pricing, and monthly runs higher. Budget for it as a second subscription, not a Claude upgrade.
A plan for each kind of writer
- Curious, writing occasionally: stay on Free. It costs nothing and the prose is the same.
- Writing daily, one person: Pro at $20/mo monthly for the first two months, then switch to the annual rate at about $17/mo. This is the default answer on this page.
- Running long agentic sessions with Claude Code: Max 5x at $100/mo, and only after Pro has genuinely blocked you more than once. See our 30-day Claude Pro review for where the ceilings land in practice.
- Team of five or more, all using Claude: Team at $20 per seat on annual billing, with usage limits confirmed with Anthropic first.
- SEO content operation: Pro plus a dedicated SEO tool. See what AI writing tools really cost for the full price ladder across every vendor in our test.
- Comparing general assistants generally: Claude vs Gemini covers the third option at this price point.
An honest note on our links
Claude has no affiliate programme, so every Claude link on this page is a plain link and there is nothing for us to earn. We recommend Pro because it produced the best prose in our own benchmark, not because it pays. The tools that do pay us - Rytr and Jasper - are marked as sponsored wherever they appear; see our affiliate disclosure. For the questions to ask before subscribing, see the buyer's checklist.
Start on the free tier, then decide
Claude Free runs the same writing model as Pro - the difference is headroom, not quality. Try it for two weeks before you spend $20/mo.
Try Claude Free →Frequently Asked Questions
Is Claude Pro worth $20/mo?
For anyone writing daily, yes. Claude recorded about 8 minutes of editing per piece in our 30-day blind benchmark - the joint-fastest in the set, level with Jasper - and the free tier's rolling caps become the limiting factor within about a week of daily use. Pro is the plan we would start on; skip it only if you need SEO tooling or templates, which Claude does not have at any price.
Is Claude really free?
Yes, there is a genuine free tier with no credit card required. It runs on a rolling usage budget with a roughly five-hour session limit, so it suits a few short drafts a day rather than sustained daily output. The strongest models, the highest limits and Claude Code sit above it.
What is the difference between Claude Max 5x and Max 20x?
Only capacity. Max 5x is $100/mo and Max 20x is $200/mo, and both are described by Anthropic as multiples of the Pro plan's usage allowance, with the 20x tier also getting the highest priority at busy times. Max does not unlock extra features, and it does not enlarge the 200K-token context window - that stays the same across Pro, Max 5x and Max 20x.
How much does Claude cost per month?
Free is $0. Pro is $20/mo on monthly billing, or about $17/mo if you pay annually ($200 for the year). Max 5x is $100/mo and Max 20x is $200/mo. Team is $25 per seat per month, or $20 per seat on annual billing. All prices verified September 2026.
Why is Claude cheaper if I subscribe in a browser?
Mobile app store purchases carry app-store commission, so the same Claude subscription bought inside the iOS or Android app is priced above the web price. Subscribe on claude.ai in a browser, then sign in to the mobile app with the same account, and you keep the cheaper web rate.
Is Claude better than ChatGPT for writing?
In our blind benchmark Claude scored highest on raw prose quality and needed about 8 minutes of editing per piece against ChatGPT's 10. ChatGPT matches the price at $20/mo for Plus but adds image generation and a much wider integration surface. If prose quality is the deciding factor, Claude wins; if you want one subscription for everything, ChatGPT is broader.