News · 2026-10-31

Claude Ultra removes the five-hour cap for verified accounts, while weekly limits remain

Anthropic’s Claude Ultra plan removes the rolling five-hour usage cap for verified accounts while retaining a weekly allowance of ten times Pro. The $200-a-month tier replaces Max 20x for existing subscribers without a price change, but its new name and usage windows have generated confusion about whether customers lost capacity.

Key facts

• Anthropic lists Ultra at $200 per month and Max at $100 per month.
• Verified Ultra accounts have no five-hour limit; the weekly allowance remains ten times Pro.
• Max retains five-hour and weekly limits, with five times Pro’s usage in each window.
• The primary source is Anthropic’s Ultra plan documentation; current terms were checked for this October 31 briefing.

The useful change is about when a customer can spend an allowance. A user doing a long coding session can hit a short rolling limit even while plenty of weekly capacity remains. Removing that shorter limit lets a verified Ultra subscriber concentrate more of the weekly allowance into a sustained period of work.

A transit pass provides a concrete analogy. Suppose it allows a fixed number of trips each week, but also restricts trips during any five-hour period. Removing the shorter restriction makes the pass more flexible. It does not increase the weekly trip total, promise a seat on every train, or make travel unlimited.

The verification condition matters. Anthropic says it uses account and payment history among its signals, and newer accounts can retain a temporary five-hour limit. In the usage display, a verified Ultra account will no longer show the “Current session” bar, according to the company’s help material. That is a practical way to check the account’s current state, rather than assuming every $200 subscriber already has identical treatment.

The historical naming issue is separate. The label Max 20x led some customers to read twenty times as a general allowance. Anthropic’s subscription lead Mike Overell explained that it described the five-hour window; the weekly cap introduced later was ten times Pro. His explanation is reproduced in Vanja Petreski’s contemporaneous account, which also discusses an Anthropic migration email. The original social post was inaccessible during the dossier’s source check, so the historical explanation should be attributed through that account.

The current pricing page and Max help page independently establish today’s terms. They do not by themselves supply a dated record of every previous allowance. The evidence supports a clearer label and a removed short-window restriction for verified Ultra accounts; it does not establish a reduction from twenty times Pro per week to ten times.

There are additional limits to the apparent simplicity. Anthropic pools usage across covered Claude experiences, including Claude Code, rather than giving each interface a wholly separate entitlement. The company’s usage-limit explanation is therefore relevant to anyone splitting work between chat and a coding assistant. The Ultra documentation also describes a separate weekly limit for Fable and reserves other possible model or feature caps.

Those details are why a multiplier should not be treated as a guaranteed number of completed jobs. A long conversation, large inputs, and demanding tasks can consume allowance differently. Readers should distinguish the entitlement a plan states from how much useful work their own workload produces before a limit is reached. The site’s inference-cost lesson explains the underlying distinction between raw computation and the cost of completing a task.

The reception evidence is narrower than some headlines suggest. The supplied Reddit feed contains posts about the rename, the missing five-hour bar, and a question about whether users were shortchanged. It contains titles and timestamps rather than measured usage or a representative sample of subscribers. A complaint can identify confusing communication without proving a quota reduction.

The strongest customer counterargument is still substantial: a product that uses several clocks, model-specific constraints, and account verification can be hard to budget around. A clearer name helps only if users can see the limits that actually apply to their account. Conversely, the strongest case for the change is practical flexibility for heavy users, even when the total weekly pool stays fixed.

The exact effective date is less secure than the current terms. October 30 announcement and migration reporting is corroborated in the dossier, while the retrieved first-party plan pages are current documents rather than a dated rollout log. For a subscription decision, the decisive facts are the current price, the account’s verification state, and the weekly ceiling. Ultra removes one interruption point for eligible users; it does not sell unlimited compute.

Primary source, verified: read the paper →

Ground Truth. (2026, October 31). Claude Ultra removes the five-hour cap for verified accounts, while weekly limits remain. Ground Truth. https://groundtruth.day/news/claude-ultra-five-hour-cap-weekly-limits.html

@misc{groundtruth:claude-ultra-five-hour-cap-weekly-limits,
  title  = {Claude Ultra removes the five-hour cap for verified accounts, while weekly limits remain},
  author = {{Ground Truth}},
  year   = {2026},
  month  = {oct},
  url    = {https://groundtruth.day/news/claude-ultra-five-hour-cap-weekly-limits.html}
}

Topics: anthropic · claude · subscriptions · coding-agents · ai-economics