Claude Pro vs Max vs Ultra: Usage Limits and Prices in India

What do Claude Pro, Max and Ultra cost in November 2026?

List prices below, then rupees at the RBI reference rate of Rs 93.8579 per US dollar for 9 November 2026. Dollar prices exclude tax. The India column shows prices TechCrunch and Business Today reported in July, which include GST and local taxes; Anthropic did not comment, and app-store prices can differ, so confirm at checkout. We could not confirm a rupee price for Ultra or premium Team seats. Enterprise ($20 a seat billed annually, 20-seat minimum, plus usage at API rates) is a separate purchase.

Is a bigger Claude plan better value?

On weekly usage, no. Pro is $20 for 1x, Max $100 for 5x and Ultra $200 for 10x: $20 per unit at every level, with annual Pro cheapest of the three at $17. Upgrading gets you more room at the same unit price. The extras are what differ. Ultra drops the five-hour window on verified accounts. Max and Ultra add $100 and $200 a month in API credits for building on the Claude Platform, so a developer who would spend that anyway effectively gets the fee back in credits. In India the sum tilts further towards Pro: Max at Rs 11,999 is Rs 2,400 per unit against Rs 2,000 for annual Pro, which is 20% more.

Is rupee billing cheaper than paying in dollars?

At this week's rate it is slightly dearer on paper. Business Today noted that dollar billing used to add 18% GST and foreign transaction fees. Annual Pro: Rs 1,596 x 1.18 = Rs 1,883 before card fees, against Rs 2,000 reported. Max: Rs 9,386 x 1.18 = Rs 11,075, against Rs 11,999. That is about 6% and 8% above dollars plus GST, in exchange for a fixed rupee bill with no foreign card charge.

How do the five-hour and weekly limits work?

Claude measures you on two clocks, and whichever runs out first stops you.

• •The five-hour window starts with your first message and resets in full five hours later.
• •The weekly limit covers all models and resets at a fixed time assigned to your account.
• •claude.ai, Claude Code and Claude Desktop share one pool, so a morning of chat comes out of the afternoon's coding.
• •Some plans add a separate weekly Fable limit inside the all-models limit. The 50% is a ceiling on Fable and adds nothing to your allowance, and Anthropic says Fable uses limits faster than other models.
• •Anthropic can change limits at its discretion to manage capacity.

Max 20x is now Ultra: why is it 10x and not 20x?

On 30 October 2026 Max 20x was renamed Ultra at the same $200, and subscribers moved automatically. Anthropic's help article says "10 times the Pro plan's usage every week", so ignore older guides that say 20x. Wokey, a third-party tracker, measures about 11x from real accounts. The missing five-hour limit applies to verified accounts only. Verification is automatic, based on account and payment history. If Settings > Usage shows no Current session bar, you are verified.

Which resets do you have, and when do they expire?

A Fable reset restores only what you spent on Fable within the weekly limit, and refills the five-hour window if you have one. The 10 November reset covers every model and limit. No Fable reset raises your limit or can be undone, and cancelling or downgrading forfeits an unused one. For anyone searching: there is no Fable 4.5. The line began with Fable 5 in June 2026.

Did Anthropic really use up more of your limit than it should have?

By its own account yes, though the size of the problem is unknown. According to Startup Fortune, the ClaudeDevs account on X said late on 10 November that Anthropic had fixed a few small issues that used more limits than intended and gave Pro, Max, Ultra and Team users one free full reset. It did not say which bugs, when they began or how much usage they cost, and said not everyone was affected. A bug report filed on 9 November by a Team-plan user described faster drain since about mid-October while logged tokens fell; that is one account and proves no cause. So do not judge your plan on recent weeks. Measure a normal week after the fix before paying to upgrade.

Why do Claude limits run out faster than you expect?

Length of conversation matters more than the length of your message. Anthropic's Claude Code docs warn that "a one-line question in a session that has been open all day still draws usage for the whole conversation" and that "one debugging session can consume more than a day of chat".

• •Resent context: each request carries the full conversation, and each tool call sends another.
• •Cache misses: on subscription usage the main conversation gets a one-hour cache. Return after a longer break and the whole context is reprocessed.
• •Subagents: each sends its own requests on a five-minute cache by default, and at xhigh and max effort Fable 5.2 can start them to review its own work.
• •Agent teams: "approximately 7x more tokens than standard sessions when teammates run in plan mode".
• •Scheduled tasks fire on their interval and send your full context each time, even while you are away.
• •Mid-task switches: changing model or toggling fast mode starts a fresh cache.

What does a long Fable session cost on usage credits?

Usage credits are billed at standard API rates, so the arithmetic is public even though Anthropic does not publish how plan limits convert into tokens. An illustrative turn with assumed sizes: 300,000 tokens of context from cache, 5,000 new input and 3,000 output on Claude Fable 5.2, at $8 input, $40 output, $0.20 cache read and $10 five-minute cache write per million.

• •Cached context: 300,000 x $0.20 / 1,000,000 = $0.06.
• •New input: 5,000 x $8 / 1,000,000 = $0.04.
• •Output: 3,000 x $40 / 1,000,000 = $0.12.
• •Total: $0.22 a turn, about Rs 21. One hundred turns: $22, about Rs 2,065.
• •One cache miss on that context: 300,000 x $10 / 1,000,000 = $3.00, about Rs 282, fifty times the cached read. On credits the main conversation drops to a five-minute cache by default, so a tea break can trigger it.
• •Pro to Max costs $80 more, about Rs 7,509, which buys roughly 363 of these turns.

Are usage credits ever worth it compared with upgrading?

Only for occasional overflow. Wokey estimates a Pro week is worth about $154 of API usage (third-party, assuming about 92% cache reads; it has not published Max). If Max gives 5x Pro, the extra 4x would be about $616 a week at API rates, for $80 more a month. Anthropic's docs put metered Claude Code at around $13 per developer per active day, or $150 to $250 a month across enterprise deployments (about Rs 14,079 to Rs 23,464), above Max's $100. If you buy credits in more than one month out of three, move up a plan.

What should you do before buying usage credits?

Work through these in order, inside Claude Code.

Which Claude plan should you buy?

Start one level lower than you think you need, and move up only after hitting a limit twice in normal weeks for reasons the routine above could not fix.

Frequently asked questions about Claude usage limits and plans

Short answers to the questions people ask most.

Where Tech Arion can help

Tech Arion builds with Claude Code and advises teams on which plan or API setup fits the work. Our AI consulting service at /services/ai-consulting can review seats against metered usage for your team.