Claude Pricing (2026): Every Plan and API Cost Explained
Claude costs $0, $20, $100 or $200 a month, but none of the plans gives you a fixed number of messages. Here's what each plan includes, how the five-hour and weekly limits work, what the API costs per million tokens and what buyers in India really pay.
Claude pricing starts at $0 for the Free plan, then $20 a month for Pro ($17 a month if you pay $200 for the year), $100 for Max and $200 for Ultra. The API is billed separately per million tokens, from $0.10 for input on Haiku 5.5 to $8 on Fable 5.2.
Here's the catch: none of the chat plans gives you a fixed number of messages. Anthropic sells usage limits that reset every five hours and every week, so a plan's value depends on how long your chats run and which model you use. For buyers in India, the dollar price isn't the bill either. A $20 Pro plan ends up near $24.58 once 18% GST and a typical 3.5% card markup are added.
Read the plan table and the API rates below, then use Claude pricing the cheap way: start with the smallest plan that covers a normal week.
Free costs nothing and includes Sonnet and Haiku with limited usage.
Pro is $20 monthly or $200 a year and adds Opus, Claude Code and Projects.
Max ($100) and Ultra ($200) are monthly only and include $100 and $200 a month in API credits.
API prices run per million tokens, and the Batch API halves them.
Claude plans and prices
Every figure below comes from Claude's pricing page, checked on 14 November 2026. Prices are in US dollars and exclude tax.
Plan
Price
Usage vs Pro
Billing
Free
$0
Limited, every 5 hours and every week
n/a
Pro
$20/month or $200/year
1x (at least 5x Free per session)
Monthly or annual
Max
$100/month
5x every 5 hours and every week
Monthly
Ultra
$200/month
10x weekly, no 5-hour limit for verified accounts
Monthly
Enterprise
$20/seat/month plus usage at API rates
Set by your organization
Annual
Team is sold for groups of 2 to 150 people, but the page text we could read didn't print a seat price, so check the live page for it.
Annual billing only exists on Pro. Paying $200 up front instead of $240 over twelve months saves $40, which is worth it only if you already know you'll use Claude all year. Our annual vs monthly billing guide has the break-even rule.
What each plan actually includes
Free includes web search, file creation, code execution, memory, connectors, Artifacts and up to five Projects, on Sonnet and Haiku.
Pro is the plan most people need. It adds Opus, Claude Code, unlimited Projects, Research, Claude in Chrome and Microsoft 365, and the option to turn on usage credits. Fable, Anthropic's top model, is available on Pro only through those credits.
Max and Ultra change three things: more usage, priority access at busy times and Fable included for up to half your weekly limit. They also come with monthly API credits for the Claude Platform, $100 on Max and $200 on Ultra.
The jump from Pro to Max is 5x the price for 5x the usage. That's fair value, but only if you keep hitting Pro's limit.
How the usage limits work
Claude doesn't count messages. Anthropic's pricing FAQ says most accounts have a limit that resets on a rolling five-hour window, paid plans add weekly limits on top, and there's no fixed message count because it depends on conversation length, complexity, the model and the features you use.
Everything draws from one pool. Chats on web, desktop and mobile share the same allowance as Claude Code in the terminal, so a long coding session eats into your chat time.
When you hit the limit you have three choices: wait for the reset, move up a plan, or turn on usage credits on a paid plan and keep working at standard API rates. Settings > Usage shows where you stand.
Long chats with big attachments burn limits fastest, so start a fresh chat for each new task. We explain how rolling limits compare with credits and tokens in our guide to AI tool pricing.
Claude API pricing per million tokens
The API is a separate, pay-as-you-go product. Anthropic's API pricing documentation, checked on 14 November 2026, lists these rates for current models:
Model
Input per MTok
Output per MTok
Fable 5.2
$8
$40
Opus 5.5
$4
$20
Opus 5
$5
$25
Sonnet 5.5
$2
$10
Sonnet 5
$2
$10
Haiku 4.5
$1
$5
Haiku 5.5 (prompts up to 100K tokens)
$0.10
$0.50
Haiku 5.5 (prompts over 100K tokens)
$0.50
$2.50
MTok means a million tokens. Output costs five times input on every model here, so long answers cost more than long prompts.
Take a hypothetical month: 1 million input tokens and 200,000 output tokens. On Sonnet 5 that's $2 plus $2, or $4. The same work on Opus 5.5 is $8, and on Haiku 4.5 it's $2.
Sonnet 5's $2 and $10 was first announced as introductory pricing. The documentation now says it's the standard price and the planned rise to $3 and $15 won't happen.
Hidden costs in the API bill
Three details in Anthropic's documentation matter more than the headline rate.
The tokenizer. Claude 4.7 and later models use a newer tokenizer that produces about 30% more tokens for the same text. A cheaper per-token rate can still mean a similar bill.
Cache and batch discounts. Reading from the prompt cache costs a tenth of the input price on most models, and the Batch API takes 50% off input and output for jobs that can wait. The two can be combined.
US-only inference. Setting US-only routing on Claude 4.6 and later models adds a 1.1x multiplier to every token. The default global routing doesn't.
Web search adds $10 per 1,000 searches, on top of the tokens it uses.
Paying for Claude from India
Claude's page shows prices before tax, and foreign software firms add 18% GST for Indian buyers without a GSTIN. Add a card's forex markup, such as SBI Card's 3.5% plus tax, and $20 becomes about $24.58. Our guide to paying for software subscriptions from India walks through the full calculation.
The annual Pro plan has one more catch. RBI rules let recurring card charges run without your approval only up to ₹15,000, and $200 is above that at any rate over ₹75 to the dollar. Expect to approve the renewal yourself, or it fails.
Use a Visa or Mastercard with international use switched on. If you run a GST-registered business, add your GSTIN in billing settings.
Which Claude plan to pick
Free: you ask a few questions a day and don't need Opus or Claude Code.
Pro monthly: you use Claude most workdays. Start here.
Pro annual: you've been on Pro for a few months and hit no surprises. It saves $40 a year.
Max: you hit Pro's five-hour limit several times a week, or you'd spend $100 a month on the API anyway.
Ultra: you work in long sessions all day and keep running into the five-hour limit on Max.
API only: you're building an app or automating a task, not chatting.
Cancelling is simple but timed. Anthropic says to cancel at least 24 hours before the renewal date, and payments are generally non-refundable outside the EEA and UK, where a 14-day withdrawal window applies.
Prices and plan details checked on 14 November 2026 from [Claude's pricing page](https://claude.com/pricing) and [Anthropic's API pricing documentation](https://docs.claude.com/en/docs/about-claude/pricing). Based on public pricing and documentation, not hands-on testing. Prices exclude tax and change often, so check the live pages before you buy.
Our take
Claude pricing is easy to read and hard to predict, because the plans sell limits, not a fixed number of messages. Most people should start on Free, move to Pro at $20 when the limit gets in the way, and only consider Max after a month of hitting Pro's cap several times a week. Annual Pro is a good deal for steady users, but in India it needs your approval at renewal.
For developers, the API is often cheaper than a subscription for light, predictable jobs. Pick the smallest model that does the task, batch what can wait and cache long prompts. Remember the newer tokenizer before comparing per-token rates with older models.
The same rule from our guide to saving money on software subscriptions applies here: pay for the usage you have, not the usage you might need. Check current Claude coupons or browse more AI tool deals before you upgrade. Get the plan size right and Claude pricing stays a fixed, sensible bill.
FAQs
The Free plan has limited usage that resets every five hours and every week. Anthropic doesn't publish a fixed message count because usage depends on conversation length, the model and the features used. Free includes Sonnet and Haiku, web search, file creation and up to five Projects, but not Opus or Claude Code.
Topics
ai tools
C
Written by
Coupo9 Editorial
Every statistic in our guides names its source, and prices come from the brand's own pricing page. When the facts behind a guide change, we update it or take it down.
Midjourney Pricing (2026): Plans, GPU Time and Which to Pick
Midjourney costs $10, $30, $60 or $120 a month, but you're really buying GPU hours. Here's what each plan includes, how many images and videos the hours cover, when Relax mode makes Standard the whole bill and why annual billing is riskier than it looks.