Loading page…
The case turns on whether a headline multiplier tied to a five-hour session remains clear when separate weekly limits govern a subscriber’s longer-run capacity.
Listen to this story
A proposed California class action could test whether Anthropic’s Claude Max labels communicate usable capacity accurately. The complaint alleges that weekly Sonnet 4 estimates imply about 3.5x Pro usage for the $100 Max 5x tier and roughly 6x for the $200 Max 20x tier—calculations based on Anthropic’s own ranges, not independent measurements or a court finding.
Karl Kahn filed the case June 14 in the Northern District of California; the proposed class covers U.S. purchases and upgrades since April 9, 2025.
Anthropic advertises 5x and 20x usage per five-hour session while separately disclosing weekly, monthly, model-specific and feature-specific caps.
The complaint cites 140–280 weekly Sonnet 4 hours for Max 5x versus 40–80 for Pro, and 240–480 for Max 20x.
Anthropic is facing a proposed class action over Claude Max plans advertised as 5x and 20x more usage than Pro. Using Anthropic’s weekly Sonnet 4 estimates, the plaintiff’s lawyers calculate roughly 3.5x weekly usage for Max 5x and 6x for Max 20x. Those are plaintiff calculations from company estimates, not independent capacity measurements or court findings.
Karl Kahn filed the proposed case against Anthropic in the U.S. District Court for the Northern District of California on June 14. He alleges that the $100 Max 5x and $200 Max 20x tiers provide substantially less weekly usage than their labels imply. The proposed class covers U.S. customers who bought or upgraded to either plan through Anthropic’s website or desktop app from April 9, 2025 onward.
Anthropic describes its Max tiers as providing 5x or 20x Pro usage per five-hour session. Its pricing page separately says paid subscriptions may carry weekly, monthly, model-specific and feature-specific caps. Kahn’s complaint argues that customers reasonably understood the multipliers to describe usable capacity over longer periods, making the weekly ceilings misleading.
Usage is not a fixed number of prompts. Anthropic says it changes with conversation length, model choice, features and task complexity. Claude Code also draws from the same pool as Claude conversations, linking sustained coding work to the allowance available in the company’s other interfaces.
The case brings California consumer-protection, false-advertising, negligent-misrepresentation, contract and unfair-competition claims. It seeks restitution, damages, an injunction and class certification. The unresolved issue is whether Anthropic’s disclosures sufficiently clarify the difference between a session allowance and longer-term capacity when customers select a premium plan. No class has been certified, and the court has not ruled the advertising false or found Anthropic liable.
Loading discussion...
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.