Claude Pro vs Max (and Ultra): what changes, and when a Token Plan is better
Claude Pro is $20, Max $100 and Ultra $200 a month. Here is what each adds, how their 5-hour and weekly limits differ, and when a Token Plan's fixed dollars of usage fit better.
Anthropic's three individual plans differ in price, in how much usage they allow, and in how that usage is rationed. Here is the difference, from Anthropic's own pages as of November 4, 2026.
Pro, Max and Ultra side by side
| Pro | Max | Ultra | |
|---|---|---|---|
| Price | $20 a month ($17 billed annually) | $100 a month | $200 a month |
| Usage | "More usage than Free" | "5x Pro usage" | "10x Pro usage every week" |
| 5-hour limit | Yes | Yes | None for verified accounts |
| Weekly limit | Yes | Yes | Yes |
| Fable | Paid with usage credits | Up to half your weekly limit | Up to half your weekly limit |
| API credits | — | $100 a month | $200 a month |
Sources: claude.com/pricing and Anthropic's Claude Code help article. The help article adds that the limits are shared across Claude and Claude Code, and that the API credits included with Max and Ultra can't be used in Claude Code or the Claude apps.
What "5×" and "10×" mean
They are multiples of Pro, and Anthropic doesn't publish Pro's allowance in tokens, messages or dollars. How fast you reach a limit depends on the model, how long your chats are and how large your files are. So the honest comparison between Pro and Max is: Max is five times whatever Pro turns out to be for your work.
When to move from Pro to Max
- You hit Pro's 5-hour limit most days, and waiting costs you more than $80 a month of time.
- You want Fable without paying for usage credits.
When a Token Plan is the better step
A Token Plan replaces the multiple with a number. $100 a month buys $300 of usage at list price, with no 5-hour or weekly windows:
| Plan | You pay / month | Usage at list price | ≈ tokens on Claude Opus 5.5 | ≈ tokens on Claude Sonnet 5.5 | ≈ tokens on GLM-5.3 |
|---|---|---|---|---|---|
| Solo | $20 | $60 | 48M | 96M | 210M |
| Pro | $100 | $300 | 240M | 480M | 1B |
| Max | $200 | $600 | 480M | 960M | 2.1B |
Token estimates assume coding-agent traffic: 85% cached input, 12% fresh input, 3% output.
It suits you if you hit the weekly limit mid-week, if you want to switch between Claude and GPT, Gemini, GLM or Kimi, or if you just want to know what your month buys. It doesn't include Fable (Claude Fable 5.1 and Claude Fable 5 are pay-as-you-go only) or the Claude apps — it is model usage for Claude Code and other tools.
And if you only use Claude, every day, using every window, Max or Ultra may give you more Claude than a plan's credit — Anthropic doesn't publish the numbers that would settle it either way.
Prices on this page come from the model catalog and were current on November 12, 2026.