Claude Code Weekly Limits Fall 17%

Anthropic is correcting a capacity promise while marketing it as an upgrade; power users will feel the downgrade immediately.

• –Usage effectively moves from 150% of baseline to 125%.
• –The change affects major paid plans, including team and enterprise seats.
• –Actual capacity varies with model choice, context length, tool usage, and effort, so prompt counts remain unpredictable.
• –Developers running long autonomous sessions, parallel agents, or frequent refactors should expect to hit weekly limits sooner.
• –Anthropic has teased better usage visibility and control, but provided no details or delivery timeline.