Weekly cap on Claude Max lands in an expanded suit

An expanded class-action lawsuit filed Tuesday alleges that Anthropic sold its $200-a-month Claude Max tier on a promise of 20 times more usage than the $20 Pro plan while a separate weekly ceiling limited what subscribers could actually run. The Verge first reported the complaint, and Thenewstack laid out how the plan's limits stack.

At a glance

• Anthropic's documentation places the 5x and 20x multipliers "per session," with limits resetting every five hours, on top of a weekly cap that applies across all models.
• Plaintiffs say the weekly limits arrived in late July 2025, months after Max launched in April, while the $100 Max 5x and $200 Max 20x tiers kept the same usage claims.
• Anthropic's motion to dismiss an earlier version of the case argued that buyers could reach the limit details through hyperlinks at purchase, comparing them to a product label a shopper can turn over.

The case reads as a test of how much an AI provider must spell out before a developer subscribes. Fixed monthly prices sit on top of compute demand that varies with every task, and a multiplier describes how two plans compare rather than how much coding fits inside either one. A ruling for the plaintiffs would likely push the sector toward more explicit disclosure of reset windows and capacity restrictions at signup.

Anthropic's multipliers reset every five hours, with a weekly cap above them

Anthropic's documentation states that the Max 5x and 20x multipliers apply per session, with usage limits resetting every five hours. The weekly cap applies across all models, and Anthropic says it may add other restrictions to manage capacity. Claude Code sessions count toward the same plan limits, so interactive coding draws on the same budget as everything else.

Once the included usage runs out, developers wait for a reset or pay more to keep working. The multiplier describes how Max compares with Pro rather than how much coding a $200 subscription buys, and workloads vary: a quick fix costs little, while a complex job burns through the allowance far faster.

The complaint dates the weekly limits to late July 2025, after an April launch

According to the complaint, which The Verge first reported, Anthropic introduced weekly limits in late July 2025, months after Max launched in April, while continuing to market the plans with the 5x and 20x usage claims. The plaintiffs allege those constraints were not adequately disclosed during subscription.

Anthropic has pushed back on that characterization. In its motion to dismiss an earlier version of the case, the company argued that customers could access information about the limits through hyperlinks during purchase, comparing those disclosures to the information on a product label that a shopper can find by turning the package over before buying.

Codex usage depends on task and model, with Astra priced 2.5x above GPT-5.6 per token

OpenAI accounts for the same variability with Codex. Its documentation tells subscribers that usage depends on the task, the model, and where the work is being run, and Codex can draw from a shared allowance with other agentic products. Once that allowance runs out, users can buy more credits.

Model choice moves the ledger as well: Astra costs 2.5x more per token than GPT-5.6, so the same allowance goes much further on one model than the other. OpenAI has also made Codex more autonomous, letting it keep working while it waits for a developer to respond. The lawsuit does not accuse OpenAI or other providers of wrongdoing.

OpenAI has been testing outcome-based pricing with some enterprise customers, charging only when an agent completes a task rather than metering raw compute. That approach carries its own problems, since someone has to define success, and failed agent runs become the provider's expense.

What the Tuesday filing leaves unsettled

Anthropic's motion to dismiss answered an earlier version of the case rather than the expanded complaint filed Tuesday, and the outcome of that filing remains open. Neither the 20x multiplier nor the five-hour reset window converts into a stated volume of work, and the published limits give no figure for what a week of Claude Code allows. Disclosure at signup is what the plaintiffs are asking the court to define.