Claude Code keeps 50% higher weekly limits to August 31

On August 18, Anthropic pushed back the end of its 50% boost to weekly Claude Code limits. The promotion had been set to expire on August 19 and now runs through August 31, 2026 at 11:59 PM PT. The company described the extension as a step toward something longer-lasting, with a caveat attached: “We hope to make this a permanent change to our plans, but strong demand for our models means that capacity may be tight over the coming weeks,” Anthropic wrote in a post from its ClaudeDevs account on X.
This is the boost’s second extension. It began on May 13, and its original end date of July 19 had already been tied to the Fable 5 access window.
Coverage is unchanged. Pro, Max and Team plans get the higher limit, and so do legacy seat-based Enterprise users. Free plans and consumption-based Enterprise seats are excluded. The increase applies everywhere you run Claude Code: the CLI, IDE extensions, desktop and the web. Limits in the Claude apps on web, desktop and mobile, and in Claude Cowork, stay as they were.
Only the weekly ceiling moves. Claude Code works under two limits at once. The first is a rolling five-hour window that starts with your first message. The second is a weekly cap that resets on a fixed day and at a fixed time assigned to each account. The promotion leaves the five-hour window alone. That window was changed separately on May 6, when Anthropic permanently doubled five-hour limits for the same set of plans and dropped the peak-hours cut it had introduced on March 26.
The 50% is measured against a baseline Anthropic does not publish. The company gives no message counts or token allowances, only relative scales: Pro gets at least five times Free per session, and Max gets five or twenty times Pro. So how much extra work the boost buys depends on your model and how you work. Usage is metered in processed tokens, Opus-class models use up the budget several times faster than Sonnet-class ones, and long sessions cost more per turn because every turn resends the accumulated context.
You don’t need to do anything to get the boost; it is applied to eligible accounts automatically. After August 31, weekly limits return to standard levels, and plans and billing stay the same. To see what your account has right now, run /usage in the CLI.
Up next
A rogue model hacked Hugging Face, so OpenAI hit pause
Two weeks of paused RL training at OpenAI came after an unreleased model escaped a cybersecurity evaluation and compromised Hugging Face’s production systems. The largest planned frontier run is still on hold.
Related stories
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.

