Claude Max Usage Claims: What the Ruling Means
Claude Max usage claims remain in court. See what the ruling says about session limits, weekly caps, subscription disclosures and AI product risks.
Claude Max's Usage Claims Are Headed to Court. Here's Why
A subscription can promise "more" in a single sitting and still fall short by the end of the month - at least, that's the gist of a proposed class action against Anthropic over its Claude Max plans, and it's enough for some claims to move forward in court.
The case focuses on Anthropic's claim that its $100 and $200 Max plans provide five or 20 times more usage than its Pro plan. The plaintiffs argue those multipliers describe usage during a five-hour session, not the total a customer gets across a week. A federal judge in California dismissed several claims but allowed allegations of negligent misrepresentation and breach of contract to proceed, according to Courthouse News. The report also says the judge wasn't convinced a help-page disclosure - buried two hyperlinks away from the main sales pages - fixed the problem.
This doesn't mean the court found that Anthropic misled customers or broke its contract. It's a decision to let particular claims continue. Still, it raises a practical question for AI companies and buyers alike: when a plan sells "more usage," what time window is it actually promising?
This article lays out the allegations, the judge's reported reasoning, what the decision does - and doesn't - mean, and what AI product teams can do to make usage limits easier to understand.
What are the Claude Max usage claims?
The proposed class action, Kahn v. Anthropic, was filed in June in federal court in Northern California. Plaintiffs Karl Kahn and Solomon Smith allege that Anthropic's description of Claude Max usage gave buyers a misleading idea of how much they could use over a longer period.
As described in the report, Anthropic's sales message said its $100 and $200 Max plans offer five or 20 times more usage than the Pro plan. The plaintiffs say those figures aren't about total weekly usage. Instead, they argue the multipliers apply to usage during a single five-hour session, while separate weekly limits cap how much customers can use over the full week.
That distinction matters because many people treat a subscription comparison as a statement about overall capacity. Even if a plan is generous within each session, it can still fall short of what a buyer expects over a week if a separate cap restricts repeated use.
The plaintiffs' weekly comparison
The plaintiffs calculated that, for one Claude model, the $200 Max plan provides six times Pro's weekly usage, while the $100 plan provides 3.5 times as much. Those are the plaintiffs' calculations as presented in the case, and they're not necessarily a complete comparison across every model, plan condition, or customer usage pattern.
This also shows why "twenty times more" is hard to interpret without a stated timeframe. That phrase could refer to a session, a day, a week, or something else. If there are both session limits and weekly caps, the session multiplier alone doesn't answer the question most buyers care about: how much can they actually use over time.
What Anthropic argued
Anthropic argued that customers should understand the increased usage as applying per session. It also pointed to a help page explaining the limits. According to Courthouse News, the judge didn't find that disclosure persuasive, noting that it was two hyperlinks away from the primary sales pages on claude.com.
So the dispute isn't just whether usage limits exist. The key issue - at least for the surviving claims - is whether the headline plan comparison, together with where the qualifying information appeared, gave prospective buyers a reasonable understanding of what they were buying.
Why did some claims survive while others were dismissed?
The reported ruling distinguishes between claims that can continue and claims the judge dismissed. Magistrate Judge Alex Tse allowed the plaintiffs' claims of negligent misrepresentation and breach of contract to proceed. Courthouse News reported the ruling, and The Next Web said it hadn't reviewed the decision itself - so the details here reflect the reporting rather than an independent read of the court document.
The judge's reasoning, as described in the report, focused partly on how far the session-based explanation was from the sales message. As Courthouse News reported, Tse wrote: "Anthropic's per session disclosure is located two hyperlinks away from the primary sales pages on claude.com." The report also says Tse found that customers paying monthly could reasonably read the usage promise as applying each month.
That matters because customers often use the billing period as the reference point for evaluating a subscription. If a plan renews monthly and advertises more usage without making the relevant measurement window stand out, a buyer may assume the promise covers the broader billing period. Whether the plaintiffs can prove their allegations is a separate issue; allowing claims to proceed isn't the same thing as deciding they're correct.
Claims that remain in the case
According to the report, the negligent misrepresentation and breach-of-contract claims can proceed. In practical terms, that means the plaintiffs can keep arguing that the usage representation was misleading and that the plans didn't deliver what the contract promised.
The ruling also treated the plaintiffs differently on a state consumer-law claim. The report says Smith's claim under that law survived, while Kahn's was dismissed but allowed him to add more detail. Dismissal with leave to amend doesn't mean the court accepted the claim as originally pleaded.
Claims the judge dismissed
Tse dismissed claims under two California laws dealing with unfair business practices and false advertising. The judge also dismissed Kahn's state consumer-law claim (again, with an opportunity to provide more detail). Separately, the judge rejected a claim that Anthropic acted unfairly by tightening usage limits during billing periods.
This mix matters. It's too neat to say the decision means "Anthropic was cleared," and it's also too strong to say "Anthropic was found to have misled customers." Some theories were dismissed; others are still in play. The surviving claims will be resolved through litigation, and the report doesn't describe a final ruling on the merits.
What does the case mean for AI subscriptions and buyers?
For customers, the case is a reminder that "more usage" isn't always a complete specification. AI subscriptions can include multiple limits at the same time: a per-session allowance, a rolling time-window cap, a weekly ceiling, model-specific restrictions, and other conditions. The limit that matters most is often the one that gets hit first.
For example, imagine a service that allows more requests during each session but also enforces a weekly cap. If you use the service heavily, you might hit the weekly cap before the extra per-session allowance helps much again. In that scenario, the session multiplier can be technically accurate and still fail to answer the more important question: what can I use across the billing period?
This also matters more for AI tools than many people expect, because "usage" isn't always intuitive. One long prompt, multiple short requests, a model with higher computational demands, and a long working session can all consume capacity differently. You might not know which limit is controlling until you bump into it. And if your work depends on the tool - coding, research, customer support - hitting a cap can interrupt a workflow, not just slow down a casual question.
What buyers should check before subscribing
Before you pick a plan, look for the unit being compared and the time period it covers. If a plan advertises a multiple, check whether it means per session, per day, per week, or per month. Then look for other caps and ask which one will likely control your actual usage.
A practical checklist looks like this:
- Identify the measurement window. Does the advertised increase apply to one session, or a longer period?
- Find every relevant cap. Look for weekly or rolling limits in addition to session limits.
- Check whether limits vary by model. A plan-level promise might not apply the same way across models.
- Consider your actual pattern of use. A plan designed for occasional bursts may not work for daily work.
- Read the cancellation and renewal terms. A monthly price doesn't automatically mean monthly usage resets in the way you assume.
If the plan page doesn't answer those questions clearly, it's worth treating that uncertainty as part of what you're buying. That's not a claim that the service is unusable - it's just a way to avoid making comparisons based on a number that might describe only one piece of the limit system.
Why the case matters beyond Anthropic
This issue applies to other AI services too, especially those that sell access in tiers. Companies need to explain not only the nominal allowance, but also how the limits interact. When a comparison spotlights one generous window while the binding cap sits somewhere else, it can create a gap between the company's internal rules and a buyer's expectations.
That gap can have real business consequences: more support requests, cancellations, and disputes - sometimes even when the company believes its fine print is accurate. Clearer descriptions can reduce misunderstandings, but they also force teams to decide which measure best reflects the product's actual value. There's a trade-off here: a simple headline is easier to scan, while a precise explanation of multiple interacting limits takes more space.
How should AI companies describe usage limits?
The most useful lesson for product and legal teams isn't "hide nothing, caveat everything." Making a sales page unreadable won't help anyone. A better approach is to make the core unit of the promise obvious where the promise is presented, and then add deeper detail nearby - without forcing buyers to reconstruct the offer from scattered documentation.
If the advantage is a larger allowance per five-hour session, say that next to the multiplier. If weekly limits can cap total use, spell out that relationship before the customer has to click through multiple pages. Ideally, a buyer should be able to understand the main controlling limit from the comparison itself, then use help material for edge cases and operational details.
A practical disclosure model
A usage-limit explanation can work in a few layers:
- State the comparison unit. Define what "five times more" or "20 times more" measures. If it depends on model or plan context, name that too.
- Name the time window. Specify whether the figure is per session, per day, or over a week. Don't leave the timeframe for inference.
- Explain interacting caps. If another cap can limit total use, say which limit is likely to apply first and how it affects the headline figure.
- Put the full rules nearby. Make the detailed explanation easy to find during the buying decision - not only after someone runs into trouble.
This doesn't mean every buyer has to understand every technical detail before hitting "Subscribe." It just means the key constraint is clear right when people compare plans and pay. That's usually more effective than relying on a footnote that's "technically correct" but easy to miss.
Test the promise against real use
Teams should also compare plan language with actual customer behavior. If many subscribers hit weekly limits even after choosing a tier based on a session multiplier, that's a sign the comparison doesn't match how customers use the service - even if the underlying math is sound.
A basic review process can include support-ticket analysis, cancellation reasons, plan-page comprehension tests, and usage data. The point isn't to eliminate limits or promise unlimited access. Limits can be necessary to manage compute costs, protect service availability, and keep plans financially sustainable. The goal is to make sure what the company measures internally lines up with what a reasonable customer thinks the plan offers.
There's still a trade-off: more detailed disclosures can reduce ambiguity but add friction. Shorter language is easier to scan, but it can hide distinctions that matter. In practice, the best solution is often a clear, qualified headline backed by a nearby explanation - rather than a vague headline paired with a distant help page.
Key takeaways
- The lawsuit challenges how Claude Max's five-times and 20-times usage claims connect to session limits and weekly caps.
- The plaintiffs' reported calculations focus on one Claude model and compare weekly usage with Pro; these are allegations, not a final finding.
- A California judge allowed negligent misrepresentation and breach-of-contract claims to proceed, while dismissing several other claims.
- The reported decision doesn't say Anthropic is liable. The case continues on the surviving claims.
- Buyers should check the time period and every cap behind a usage multiplier. AI companies should make the measurement window clear alongside the plan claim.
Conclusion
The Claude Max case is about the gap between a usage number and a usage promise. A five-hour session allowance may be one way to measure capacity, while a weekly cap determines what a customer can sustain across repeated sessions. If plan language doesn't make that distinction clear, buyers may understand the offer differently from what the company intends.
The court hasn't resolved whether Anthropic's claims were misleading. It has allowed some claims to continue and dismissed others, leaving the central dispute open. For customers, the practical step is to check the time window and the caps behind any headline multiplier. For AI product teams, the goal is straightforward: explain the limit that actually governs use where customers make the purchase decision.