Anthropic Must Face Lawsuit Over Claude Max Usage Limits

A federal judge has ruled that Anthropic must face claims that it misled customers about Claude Max usage limits, allowing the core of a proposed class action to proceed. U.S. Magistrate Judge Alex Tse found the company's "5x or 20x more usage than Pro" advertising "plausibly misleading," Courthouse News reported. The ruling, issued in the Northern District of California, is a setback for Anthropic as it prepares for a public listing.

At the heart of the case is what "more usage" actually means. Anthropic's site promises Max subscribers "5x more usage than Pro" or "20x more usage than Pro," but plaintiffs say the higher multipliers apply only within a five‑hour session, not to the weekly caps that actually limit heavy users. "Claude Max subscribers receive five or 20 times more usage per five‑hour session than Pro users, but remain subject to weekly caps that are not five or 20 times higher," ² summarized from the lawsuit.

Anthropic argued that customers should understand "more usage" to mean per session, pointing to a help page. Judge Tse was not persuaded at this stage. "Anthropic's per session disclosure is located two hyperlinks away from the primary sales pages on claude.com," he wrote, and customers did not need to visit that page to sign up, according to Courthouse News. Plaintiffs likened the disclosure to "fine print on the back of a shampoo bottle."

The suit, first reported in June by the Wall Street Journal, targets Anthropic's premium tiers: Max 5x at $100 a month and Max 20x at $200 a month, both sitting above the $17 to $20 Pro plan, according to Quartz. Plaintiff Karl Kahn, a Washington, D.C. subscriber, said a single five‑hour coding session consumed 15% of his weekly allotment on the Max 20x tier, forcing him to halt his work or buy more usage.

The ruling was a partial win for plaintiffs. Negligent misrepresentation and breach‑of‑contract claims survive, as does a California consumer‑protection claim from plaintiff Solomon Smith, a computer‑science student who used Claude for coursework and coding projects, Reuters reported. But the judge dismissed unfair‑competition and false‑advertising claims, and rejected the argument that Anthropic tightened limits mid‑billing as a "bait‑and‑switch." "Our most important claims proceed," plaintiff lawyer Glenn Danas told,² "with a straightforward path to amendment for the claims that were dismissed."

The case is an early test of how AI companies must disclose subscription limits, and it matters most to the heaviest users. If the "5x or 20x" framing is ultimately found misleading, Anthropic and other AI vendors may have to publish exact weekly or monthly caps up front rather than burying them two hyperlinks deep. For builders who run Claude Code and other tools at the edge of their allowance, that could mean finally knowing what they are actually paying for.