Anthropic is dressing up a usage cut as a generous increase, and the math doesn’t lie. Starting September 14, the baseline weekly limits for Claude Code across Pro, Max, Team, and Enterprise plans will permanently rise by 25 percent. Sounds good — until you realize a temporary 50 percent boost has been running for a while now. Going from a 50 percent uplift to a 25 percent one means users actually lose about 17 percent of the capacity they’ve been enjoying. The official Claude account confirmed the change on X, though the framing was notably careful.
The timing is awkward. Anthropic has been aggressively courting developers with Claude Code, shipping performance optimizations and an automatic feedback tool that generates bug reports when something fails, letting users review and submit them with optional conversation logs. That kind of quality-of-life improvement suggests a product team that wants to keep developers happy. Cutting effective limits in the same breath sends a different signal: capacity is constrained, and the temporary boost was never sustainable.
What makes this sting more is the lack of hard numbers. Anthropic says it’s working on changes that will “make it feel like you’re getting more from Claude” and give users more control and transparency over usage. That’s a promise, not a policy. If the goal is transparency, the company hasn’t shipped it yet. Users are left to infer their actual limits from an ambiguous percentage change.
This isn’t the first time an AI company has played the percentage shell game. OpenAI has shuffled rate limits on ChatGPT and API access in similarly opaque ways, often when demand outpaced compute. The pattern is familiar: raise the baseline, remove a temporary boost, and hope the headline number distracts from the net loss. For developers running Claude Code daily, the only number that matters is how many prompts they can fire before hitting a wall — and that number is going down. The bug report feature is nice, but it doesn’t compile code or ship features. When capacity shrinks, so does productivity. The real question is whether Anthropic can actually deliver on its transparency pledge before frustrated developers start pricing out alternatives.
