Claude's $200 Plan Is Now Called Ultra and Drops the 5-Hour Limit for Verified Accounts. Nobody Has Defined 'Verified.' Here's the Two-Week Test Before You Upgrade.
On October 30, Anthropic renamed its two top subscription tiers. The $100 a month Max 5x plan is now just Max, and the $200 a month Max 20x plan is now Ultra. Prices did not change. What did change, according to reports of the update, is that verified Ultra accounts no longer face the 5-hour usage window, and the company has not said what "verified" means.
If you run your business on Claude Code, that last sentence is the whole story.
What changed, as far as anyone can confirm
I could not load Anthropic's pricing page myself, so the details below come from third-party readings of it, including explainx.ai, and from trade coverage of the announcement. Treat the specifics as reported, not confirmed.
The reported structure: Pro is $20 a month, Max is $100, Ultra is $200. Max and Pro keep a rolling 5-hour session window. Ultra is listed as "No 5-hour limit" for verified accounts. Weekly limits still exist on every paid plan, and the pricing page reportedly lists Ultra at 10x Pro and Max at 5x Pro for the weekly pool. Ultra also reportedly comes with $200 a month of Claude Platform API credit, and Max with $100.
Two things are unresolved. The pricing page does not define "verified," and reports say the 5-hour limit is being removed in stages, starting with existing subscribers. One report says it began for existing verified subscribers on October 30. No dates are published, and no opt-in process is described.
Why the 5-hour window mattered and the weekly cap matters more
The 5-hour window is the limit you feel. You are mid-refactor, the agent stops, and you wait for the window to reset. Removing it changes how a heavy day feels.
It does not change how many tokens you get in a week. Usage across web, desktop, mobile, and Claude Code draws from one pool, and the weekly limit sits on top of the window. Anthropic's FAQ, as quoted in coverage, also says it may limit usage in other ways at its discretion. So the most likely change for a heavy user is not "I can do more." It is "I can spend my week in three days instead of seven, and then hit the wall."
One more caveat from the same reading: the reports could not compare the new Ultra weekly allowance with the old Max 20x allowance, because the page does not state one. If you are a current 20x subscriber, you don't know yet whether your weekly pool changed.
The two-week test
I would not upgrade on a rename. I'd run this instead.
For two weeks, write down three numbers at the end of each workday: how many times you hit the session limit, roughly how much of your weekly meter is used, and what the hit cost you in lost time. You don't need precision. A note in a text file is fine.
Then apply a rule. If you hit the 5-hour wall more than twice a week and your weekly meter is under 60% by Friday, the window was your bottleneck, and a plan without it is worth paying for. If your weekly meter is already above 90% by Wednesday, the window was never the problem, and Ultra's higher weekly pool is the only part you are buying. In that case compare the extra $100 a month against what the same money buys in API usage at published per-token rates.
Usage credits are the other option. After a limit, you can turn on credits billed at API rates. Coverage lists Opus 5.5 at $4 input and $20 output per million tokens, and Fable 5.1 at $10 and $50. For a solo operator with a bursty month, paying API rates for the spike can beat a permanent upgrade.
Where this take could be wrong
If the rollout is wide and the weekly pool is generous, removing the window is a real quality-of-life win, and my caution looks like overthinking. A plan is also worth more than its token math if it keeps you in flow. I'm weighing a hard-to-measure benefit (fewer interruptions) against a measurable cost ($100 a month more than Max), and reasonable people will price the interruptions higher than I do.
The second risk goes the other way. Anthropic has changed what a plan includes before, and so has OpenAI. OpenAI recently canceled a cut to its Pro 200 tier and doubled Codex limits, which shows how quickly limits move in both directions. A workflow tuned to the exact shape of a limit breaks when the limit moves.
What I'd actually do
Stay on the plan you have this week. Start the two-week log today. Keep an API key funded with a small balance so a sudden limit does not stop paid client work. If at the end of two weeks the 5-hour wall cost you more than about $100 worth of your own time, upgrade, and set a calendar reminder to re-check the limits a month later.
Author
Lukas
@lukcombinator