Anthropic accused of overselling Claude subscriptions
A proposed class action lawsuit accuses Anthropic of misrepresenting how much usage Claude subscribers actually receive, alleging that the attention-grabbing multipliers attached to its paid plans do not reflect what customers get in practice. The case, first reported by The Verge, focuses on the Max tier at the top of Anthropic's consumer subscription lineup and on the distance between the figures used to sell that tier and the caps that govern it.
At the heart of the complaint is how the plan's value is framed. According to the plaintiffs, subscribers who pay $100 per month are promised five times the usage available on the Pro plan, while those paying $200 per month are promised twenty times that usage. The plaintiffs contend those advertised limits do not hold across the board, and that the actual ceiling is structured in a way the marketing does not make clear.
What the advertised multipliers are supposed to mean
The lawsuit's central allegation is that the multipliers apply only inside rolling five-hour windows and are then squeezed further by a separate weekly cap. Under that arrangement, a subscriber who reads "five times the usage" as a straightforward multiple of what a cheaper plan delivers would be working from an assumption the complaint says the company never corrects at the point of sale.
The plaintiffs argue the result is a service that, in real-world use, delivers considerably less than buyers anticipated. Their reading of the tiers looks roughly like this:
- The Max plan at $100 per month is advertised as five times the usage of the Pro plan.
- The Max plan at $200 per month is advertised as twenty times the usage of the Pro plan.
- Those multipliers are said to operate only within five-hour windows rather than as continuous allowances.
- A weekly limit sits on top of the window, further constraining how much a subscriber can consume.
Because the multipliers are expressed relative to another plan rather than as absolute quantities of usage, the complaint suggests, the headline number tells a buyer very little about what a subscription will actually support over a month of heavy work.
Anthropic says the structure is documented
Anthropic does not appear to dispute the mechanics described in the filing. The company confirms the window-based structure and the weekly ceiling on its help pages, and it reserves the right to restrict usage further at its own discretion. That discretion is itself part of what the plaintiffs point to, since a customer cannot plan around a limit that the provider can tighten unilaterally.
From Anthropic's perspective, the distinction matters: the limits are published, they are explained in support documentation, and the company retains flexibility to manage capacity across a product that depends on expensive computing resources. The argument is essentially that nothing was hidden, even if the framing emphasized the multiplier rather than the conditions attached to it.
The dismissal motion turns on hyperlinks at checkout
Anthropic has filed a motion to dismiss, arguing that all the relevant details were available through hyperlinks during the purchase process. In other words, the company contends that a subscriber who wanted to understand the five-hour windows and the weekly cap could have found that information before paying.
The plaintiffs' lawyers pushed back on precisely that reasoning, arguing that consumers have no practical way to verify what an AI service actually delivers and therefore have to rely on honest advertising. A link buried in a checkout flow, they suggest, is not the same as a clear statement of what a subscription buys. The dispute thus becomes a question about where disclosure is placed and how prominent it must be when the headline promise is a simple multiple.
A refiled case from the same firm
This is not the first attempt to bring the claim. The law firm behind the suit had already filed an earlier version of the class action earlier in the summer, making the current filing a renewed effort rather than a fresh allegation. The refiling indicates that the lawyers involved have adjusted their pleadings while holding to the same core theory about how the subscription tiers were marketed.
The identity of the firm, VacaDaffanLaw, and the reporting from The Verge are the two sources anchoring public understanding of the case so far. Because the matter is at the motion-to-dismiss stage, no finding has been made about whether the marketing was in fact deceptive.
Why usage claims are hard to verify before buying
The plaintiffs' argument about verification cuts to a broader difficulty in software subscriptions generally. A customer evaluating a streaming service or a storage plan can often reason about the units involved. An AI subscription is different: what a subscriber receives is measured in an allowance whose usefulness depends on how hard the models are working, and the provider controls both the meter and the ceiling.
That asymmetry is central to the complaint. If a buyer cannot test the service meaningfully before committing, and if the same company that sets the limits also reports them, then the advertised multiplier becomes the primary signal available at the moment of purchase. The plaintiffs' position is that this signal was made to look simpler and more generous than the underlying policy allows.
What the case signals for AI subscription pricing
Whatever the outcome, the lawsuit highlights a tension that is likely to follow other AI companies as they build consumer tiers around compute-heavy products. Marketing tends to favor clean, comparable numbers, while the operational reality involves rolling windows, weekly thresholds, and discretionary adjustments when demand surges. Bridging those two languages is where disputes of this kind tend to originate.
The litigation also puts a spotlight on how AI firms describe value. A multiplier is meaningful only if the base it multiplies is understood, and the plaintiffs argue that base was never made clear enough. If the court allows the case to proceed, the question of what a reasonable subscriber would have understood from the pricing page could become the deciding issue.
For now, the matter sits in the early procedural stage. Anthropic is seeking dismissal on disclosure grounds; the plaintiffs are seeking to have the claim heard on its merits. The outcome may influence how prominently AI providers spell out usage windows and caps in the future, and how much weight a single advertised figure can carry in a market where the product itself is difficult to inspect before purchase.
This article is based on reporting by The Decoder. Read the original article.
Originally published on the-decoder.com
