Alex Getmanearly AI signals

Analysis · Claude plans · 3 min read

Claude plans and limits: Pro, Max and Ultra

Current terms beside a calculation you can apply to your own workload.

Alex GetmanPublished 1 Nov 2026

Choose a Claude plan by the interruption it needs to remove. A short session limit, a weekly allowance and an API budget solve different problems. Here is a dated comparison and a worksheet for deciding what extra capacity would be worth to you. Prices checked on 1 November 2026.

Current individual plans

Claude's pricing page lists these USD web prices before applicable tax. Annual Pro costs $200 paid up front; monthly Pro costs $20. The table uses monthly billing throughout.

PlanMonthly priceFive-hour allowanceWeekly allowance
Free$0LimitedLimited
Pro$20BaseBase
Max$1005× Pro5× Pro
Ultra$200No limit for verified accounts10× Pro

These allowances are not fixed message counts. Anthropic's Ultra documentation confirms the verification condition and the continuing weekly limit. Removing the five-hour limit does not remove all limits.

Identify the actual bottleneck

Anthropic distinguishes usage limits from conversation length. Claude chat, Desktop and Claude Code share usage. The account's Settings → Usage shows the session and weekly reset times. A context limit concerns one conversation, rather than the amount of work allowed over a week.

Record which limit interrupts you and the task you were trying to finish. Buying capacity for a different bottleneck may leave the interruption unchanged. For a week, keep one row per interruption: time, model, task, displayed limit, and whether the work actually had to wait.

Put a value on uninterrupted work

Use the difference in price as a hurdle, then test whether the upgrade removes your actual interruption.

Monthly moveAdditional costTime worth $20/hour needed to cover it
Pro → Max$804 hours
Max → Ultra$1005 hours

The $20/hour figure is an illustrative value for your time, not a measured saving or a claim about earnings. Replace it with your own number. Four hours of blocked tasks are not automatically four hours recovered: you may have useful work to do while waiting, or the new plan may encounter another limit.

Before upgrading, write down one testable expectation, such as completing a particular daily workflow without a session interruption. Review the outcome after comparable work. Count completed deliverables and remaining interruptions, not just the size of the allowance.

Keep API capacity separate

The pricing page currently includes $100/month of API credits with Max and $200 with Ultra. Follow the linked credit terms before budgeting: a credit balance and subscription usage are different quantities. For token-cost arithmetic, use the API cost worksheet; it includes cache writes, reads and retries.

For repository work, the Claude Code setup and billing guide explains where to start and what to inspect in your account. If you are choosing between coding tools, use the Codex, Claude Code and Cursor comparison to compare subscriptions, separate usage charges and review time.

Use the Claude timeline for dated changes and the model catalogue for model-specific facts. A product announcement explains what changed on that date; this reference is the place to compare the currently checked terms.

Update the decision when the offer changes

Recheck the provider links before paying. Update prices, conditions and the review date together. Keep the interruption log after a plan change so the next decision rests on your workload rather than a larger advertised multiplier.