Anthropic Sued Over Claude Max, Plaintiff Says the $200 Plan Delivers Less Than Advertised

Anthropic has a new headache, and this one is about the bill. A proposed class action filed Monday in California accuses the company of overselling its premium Claude Max plans, claiming the 100 and 200 dollar tiers deliver far less usage than the 5x and 20x they advertise. The complaint comes from a single power user, and it lands while Anthropic is already busy fighting Washington over its models.

Here is how the plans are pitched. Claude Pro runs around 17 to 20 dollars a month. Above it sit Max 5x at 100 dollars and Max 20x at 200, sold on the promise of five or twenty times the Pro plan's compute access. The lawsuit, filed in the Northern District of California for Washington DC user Karl Kahn, says the real limits land far below that, and that they are nearly impossible for a customer to actually measure.

The mechanics are the heart of it. Usage resets on rolling five-hour windows, then a separate weekly cap limits total activity on top of that. Anthropic added those weekly caps in late August 2025, after heavy users ran Claude Code almost nonstop. Kahn's example is the headline number: one five-hour coding session ate 15 percent of his weekly quota on the top tier. At that rate, the suit argues, the 20x label does not hold up. The case seeks class-action status for buyers going back to last year.

There is a real tension underneath the legal language. AI inference is expensive, and a small number of power users running coding agents around the clock can burn compute far faster than typical subscribers. The weekly caps exist because of exactly that. So the dispute is partly about disclosure, whether Anthropic made the true limits clear, and partly about whether selling a plan as 20x sets an expectation that usage-based reality cannot meet. Anthropic has not detailed a response to the claims yet.

For Anthropic, the timing is awkward, but the stakes are mostly reputational for now. It is a proposed class action, unproven, and these cases often take a long time to go anywhere. Still, it pokes at a sensitive spot for the whole industry: opaque, shifting usage limits on premium AI plans that customers struggle to verify. With Anthropic heading toward an IPO and leading on business adoption, how it handles paying users is part of the story investors will watch.

The government fight is about whether Anthropic's models are too powerful. This one is about whether its top plan gives you enough. Different problem, same week. For subscribers, the useful takeaway is simpler: read the usage limits, because someone just went to court over them.