Claude AI Pricing 2026: Free, Pro, Max & Subscription Cost in India

Quick Answer: What Does Claude AI Cost?

Claude AI costs nothing on the Free plan. Pro is the standard individual subscription at twenty dollars a month, or two hundred dollars billed up front for a year, which Anthropic's pricing page rounds to seventeen dollars a month. Above that, Max 5x is one hundred dollars a month and Max 20x is two hundred dollars a month, both sold monthly only. Team seats are twenty-five dollars per seat monthly, dropping to twenty dollars per seat on annual billing, with premium seats at one hundred and twenty-five dollars monthly or one hundred annually. Enterprise is twenty dollars a seat plus usage billed at API rates and is annual only. In India, the same plans are published in rupees and include GST. None of these subscriptions include API access — the Claude API has its own per-token price card starting at one dollar per million input tokens.

Current Claude Subscription Pricing

The table below reproduces Anthropic's published individual and team prices as verified on the date shown at the end of this article. Two structural details matter more than the headline numbers. Max has no annual option, so there is no way to prepay it for a discount. Enterprise is annual only and is sold with a seat floor — twenty seats when bought self-serve, more when handled by sales. All prices are exclusive of applicable tax unless the row says otherwise.

Claude AI Subscription Cost in India

Anthropic began publishing India-specific rupee prices in July 2026, so Indian customers no longer have to route an international card in US dollars. Reported figures show Claude Pro at about Rs 1,999 per month on an annual plan, roughly Rs 24,000 for the year, rising to about Rs 2,399 per month when billed monthly. Claude Max 5x is around Rs 11,999 per month and Max 20x around Rs 23,999 per month. Team pricing has been reported in the region of Rs 2,399 per seat per month. Three qualifiers belong next to every rupee figure. The India prices are inclusive of GST, while the published US prices exclude tax, so the two lists are not directly comparable and we deliberately do not convert between them. Payments run through credit or debit cards and through Apple and Google app-store billing; UPI is not yet supported. And the rollout has appeared progressively, so the exact amount you see can depend on when your account and region were enabled — confirm on claude.com with your country selected before purchasing.

Claude Free vs Claude Pro

Free gives you the whole chat interface with session-based limits; Pro buys capacity and the workspaces that heavy users live in. Pro adds Claude Code, the Cowork, Design and Science environments, unlimited projects for organising chats and documents, Research, access to a wider set of Claude models, Excel and PowerPoint integration, and Microsoft 365 connectivity. Free remains the right answer for occasional use, and it is where everyone starts. Pro is the tier most individuals settle on, because the difference shows up as fewer interruptions rather than as new buttons — you keep the same interface and simply get further before hitting a five-hour window.

Pro vs Max: Who Actually Needs Max

Max is not a different model tier, it is a larger bucket. Max 5x delivers five times the Pro usage per five-hour session, Max 20x delivers twenty times, and both add higher output limits, early access to new features, and priority handling when traffic is heavy. Anthropic sells Max monthly only, and upgrading between tiers is charged on a prorated basis for the remainder of the billing cycle. The honest test is whether Pro's limits interrupt your working day. If you are a Claude Code user running long agent sessions, or you work through large document sets continuously, Max pays for itself by removing the wait. If you chat for an hour or two, Max 5x will mostly be unused capacity.

Claude Team and Enterprise Pricing

Team covers workspaces of roughly two to a hundred and fifty people and comes in two seat types that can be mixed inside one workspace. A standard seat costs twenty-five dollars per seat per month on monthly billing or twenty dollars on annual, and includes everything in Pro plus centralised billing, administration, single sign-on and domain capture, and exclusion of client data from model training by default. A premium seat multiplies usage about five times over a standard seat at one hundred and twenty-five dollars monthly or one hundred dollars annual per seat. Enterprise starts at twenty dollars per seat per month plus usage at API rates, is annual only, and adds SCIM provisioning, audit logs, custom data retention and role-based access control.

Monthly vs Annual Billing

Annual billing exists for Pro and Team, and Enterprise is annual only. Pro annual is two hundred dollars paid once, against two hundred and forty dollars across twelve monthly payments — a saving of forty dollars, or about seventeen percent. Team standard seats save fifty dollars a seat per year going annual, and premium seats save three hundred. The trade-off is commitment: an annual plan is charged up front and is awkward to downgrade mid-term, while monthly billing can be cancelled at the end of any cycle. Max deliberately has no annual plan, so there is no discount to weigh there.

Claude Subscription vs Claude API

These are separate accounting systems and neither includes the other. A claude.ai subscription is a flat monthly fee that covers a usage allowance inside the apps, and it grants no API credit. An API key is a metered billing account in the Anthropic Console that charges per token, and it grants no claude Pro access. The break-even runs one way: for sustained conversational or agentic use, a twenty-dollar Pro plan is dramatically cheaper than the API cost of the same volume, while for bursty programmatic calls the API avoids paying for capacity you never use. Teams commonly run both — seats for people, an API budget for products — and track them separately in the Console.

Claude API Pricing per Million Tokens

API prices are quoted per million input tokens and per million output tokens, with output consistently the more expensive side. Prompt caching cuts the cost of reused input by roughly ninety percent for cache hits, and the Batch API applies a flat fifty percent discount for workloads that can wait. Context windows differ by model: Haiku 4.5 is a two-hundred-thousand-token model, while the Sonnet, Opus and Fable tiers in the table carry a one-million-token window. One timing note matters in 2026 — Sonnet 5 carried an introductory two-dollar and ten-dollar rate through August 31, 2026, and now bills at the standard three and fifteen dollar rates shown below. Always confirm the live rate card in Anthropic's documentation before you model a budget.

Claude Cost Examples

Worked numbers make the price card concrete. Pro on annual billing costs two hundred dollars for the year versus two hundred and forty paid monthly, saving forty dollars. On the API, ten million input tokens and two million output tokens on Haiku 4.5 come to ten dollars plus ten dollars, twenty dollars in total; the same volume on Sonnet 5 costs thirty plus thirty, sixty dollars; on Opus 4.8 it is fifty plus fifty, one hundred dollars. A single long document conversation in the app might consume a substantial slice of a five-hour allowance but still cost nothing extra on Pro, which is exactly why subscriptions suit interactive work and per-token pricing suits products. For context on what equivalent open-weight models cost to run, our model database lists hosting and inference costs alongside these API rates.

Important Pricing Caveats

Read these before you commit. Published prices exclude applicable tax unless stated — India's rupee prices are the exception and include GST. Mobile and app-store billing can be priced differently from the web, and Anthropic notes as much for Max. Promotional offers exist and expire: extended allowance promotions and free access to newer models have run during 2026 on fixed end dates, so a price you saw during a promotion may not be the standing price. Enterprise seat floors mean the headline twenty-dollar seat price only applies above the minimum order. And Anthropic states plainly that prices and plans are subject to change at its discretion, which is why every figure on this page carries a verification date.

Limitations and Assumptions

The dollar figures come from Anthropic's own plan pages and help centre; the rupee figures come from Indian press reporting of the July 2026 rollout and from the amounts shown in-region, so treat them as accurate to that reporting rather than as a locked quote. We do not convert currencies, because the two price lists are taxed differently. Message counts are not part of pricing — Anthropic publishes relative multipliers such as five times Pro rather than guaranteed message totals, so no plan on this page should be read as promising a specific number of messages.

Frequently Asked Questions

Sources

• Anthropic — Claude plans and pricing (official published prices)
• Claude Help Center — Max plan pricing tiers and session limits
• Anthropic docs — Claude API pricing per million tokens
• India Today, 13 July 2026 — India-specific Claude pricing (third-party reporting)
• Economic Times, 14 July 2026 — rupee pricing rollout and payment methods (third-party reporting)