Anthropic said it will introduce weekly usage caps on its Claude Pro and Claude Max subscriptions starting 28 Aug. 2025, tightening control over how much customers can run the company’s generative-AI models in a seven-day period. The San Francisco-based developer explained that a small cohort of power users have been operating Claude Code continuously, in some cases running up model costs running into "tens of thousands of dollars" a month and reselling access. The new policy adds two rolling seven-day limits—one on overall usage and another specific to the company’s flagship Claude Opus 4 model—while keeping the existing five-hour caps. Anthropic estimates that fewer than 5 percent of current subscribers will hit the new thresholds. Users on the $20-per-month Pro tier and the $100 or $200 Max tiers who exceed the allowance can still purchase additional capacity at standard API rates, the company said. With AI providers struggling to secure sufficient computing resources, several rivals have also tightened allowances this year. Anthropic said the measures should stabilise service in the near term while it explores longer-running options for heavy-duty workloads.
Anthropic unveils new rate limits to curb Claude Code power users: https://t.co/WprUNDXjRP by TechCrunch #infosec #cybersecurity #technology #news
Anthropic unveils new rate limits to curb Claude Code power users | TechCrunch https://t.co/InU5kugISk
Anthropic: Rolling Out New Weekly Rate Limits For Claude Pro And Max In Late August