A Judge Let the Claude Max Usage Case Go Forward. Price Your Business on Tokens You Can Count, Not on a Plan Multiplier.
On November 7, The Next Web reported that a federal magistrate judge in Northern California let key claims in Kahn v. Anthropic proceed: negligent misrepresentation, breach of contract and one consumer-protection claim. The case is about whether Claude Max "5x" and "20x" plans delivered what the names promised. No class has been certified and nothing is decided on the merits. But if your product margin quietly depends on a subscription multiplier, this is the week to look at that.
What the case is about
Max launched at $100 a month (5x) and $200 a month (20x), pitched as five or twenty times Pro usage. The complaint says the multiplier applied per five-hour session, not to the weekly allowance, so real weekly delivery was far below the headline. The plaintiffs calculate roughly 3.5x Pro for the $100 plan and about 6x for the $200 plan on one model.
Anthropic argued customers should read the multiplier as per session and pointed to a help page. According to the report, the judge was not persuaded, noting the per-session disclosure sat two hyperlinks away from the main sales pages. Some claims were dismissed, including two California business-practice laws and a claim about tightening limits mid-billing-period.
I am reporting this from TNW's coverage, which cites Courthouse News. I did not read the ruling itself, so treat the details as secondhand.
What changed on November 1
Anthropic renamed the tiers on November 1: Max 5x became "Max" at $100, Max 20x became "Ultra" at $200. Prices are unchanged. Per TNW, the pricing page now lists Max as 5x Pro usage every five hours and every week, and Ultra as 10x Pro weekly, with no five-hour limit for verified accounts. TNW notes the announcement did not mention the shift from 20x to 10x. Whether that is a disclosure gap or a clearer label is a judgment call, and the court, not me, will weigh it.
Why this is a business problem, not a legal one
If you run a product where an agent does work on your Max subscription and you charge customers a flat fee against it, your cost of goods is a number you cannot see. A multiplier is a marketing unit. It does not tell you tokens per day, and it can be redefined by a rename.
The fix is boring. Measure your own consumption in tokens or requests per task for two weeks. Divide your monthly plan price by the tasks you actually completed. That is your real unit cost, and it is the number to price against. If you build on subscription quota rather than API billing, add a line to your model: what happens to margin if the quota halves.
The honest take
I would not read this ruling as a verdict that Anthropic misled anyone. A motion-to-dismiss ruling means the claims are plausible enough to continue, which is a low bar. The company may well win. The multiplier may also be documented clearly on the help page that the judge found hard to reach.
My recommendation does not depend on who wins. Subscriptions with opaque limits are fine for personal use and risky as a cost basis. For anything customers pay you for, I would use API billing, track cost per completed task, and keep a second model provider wired in behind a thin layer, so a quota change is a config edit and not a crisis.
Author
Lukas
@lukcombinator