Anthropic is facing a class-action lawsuit that says its Claude Max subscriptions did not deliver the usage limits the company promised. The complaint, filed in federal court in Northern California on 14 June, argues that customers paid for bigger quotas on Claude Max 5x and Claude Max 20x but hit limits far sooner than the marketing suggested.
The case could cover U.S. users who signed up for those plans, and it lands in a part of the AI business that is getting harder to explain as the bills get larger. Subscription tiers with ”more usage” are now a standard pitch across the sector, but the actual metering behind them is often opaque enough to make a lawyer smile and a customer reach for the credit card again.
What Karl Kahn says happened
The named plaintiff, Karl Kahn, says he started with Claude Pro in June 2025, then moved to Claude Max 5x in January and later to Claude Max 20x in April. Even after paying more, he says he kept running into limits and had to buy extra access once his allowance ran out.
His complaint also says Anthropic never clearly explained how ”usage” is measured on the Max plans. That omission is the heart of the lawsuit: if customers cannot tell what is being counted, they cannot tell whether they are getting what they paid for.
Claude Max pricing and claimed limits
- Claude Max 5x: about $100 a month, with 5 times more usage than the Pro plan.
- Claude Max 20x: about $200 a month, with up to 20 times higher limits.
- Claude Max 20x also includes priority access to new models and expanded workflows, including tools such as Claude Code.
The lawsuit says those promises are misleading because the real caps depend on hidden consumption metrics. That is a familiar complaint in AI, where ”more access” can mean anything from more messages to more tokens to whatever internal meter the company decides to expose, which is fantastic if you sell the plan and less fantastic if you buy it.
A familiar subscription fight in AI
Anthropic is not the only company selling premium AI access with tiered limits. OpenAI, Google, and Perplexity AI all use versions of the same model: higher prices, higher quotas, and some combination of priority access or better tools.
That makes this case more than a dispute over one brand’s wording. If the court treats opaque usage limits as a consumer-protection problem, the pressure will spread to the rest of the subscription AI market, where ”unlimited” has always had a suspiciously finite feel.
What the lawsuit wants
The complaint seeks damages, refunds, an order stopping the current way the plans are described, and class certification for users who bought Claude Max 5x or Claude Max 20x through Claude.com or the desktop app starting in April 2025. The plaintiff’s legal team says the case fits ordinary consumer-law rules against deceptive or insufficiently clear advertising.
That last point may be the most awkward for Anthropic. AI companies are racing to sell bigger plans as models get more capable and more expensive to run, but the more complex the meter, the easier it is for customers to feel cheated. Expect this lawsuit to test how much vagueness subscription AI can get away with before the fine print stops doing the heavy lifting.